Skip to main content

Full text of "HMOs, issues and alternatives for Medicare and Medicaid"

See other formats

HMOs: Issues 

and Alternatives 

for Medicare and Medicaid 






Published by the Health Care Financing Administration 
Office of Research, Demonstrations, and Statistics 

Health Care 



The Health Care Financing Administration was established to 
combine health financing and quality assurance programs into 
a single agency. HCFA is responsible for the Medicare pro- 
gram, Federal participation in the Medicaid program, the Pro- 
fessional Standards Review Organization program, and a varie- 
ty of other health care quality assurance programs. 

The mission of the Health Care Financing Administration is to 
promote the timely delivery of appropriate, quality health care 
to its beneficiaries— approximately 47 million of the nation's 
aged, disabled, and poor. The Agency must also ensure that 
program beneficiaries are aware of the services for which they 
are eligible, that those services are accessible and of high 
quality, and that Agency policies and actions promote efficien- 
cy and quality within the total health care delivery system. 

HCFA's Office of Research, Demonstrations, and Statistics 
(ORDS) conducts studies and projects that demonstrate and 
evaluate optional reimbursement, coverage, eligibility, and 
management alternatives to the present Federal programs. 
ORDS also assesses the impact of HCFA programs on health 
care costs, program expenditures, beneficiary access to ser- 
vices, health care providers, and the health care industry. In 
addition, ORDS monitors national health care expenditures 
and prices and provides actuarial analyses on the costs of cur- 
rent HCFA programs as well as the impact of possible 
legislative or administrative changes in the programs. 

^iis HMOs: Issues and Alternatives 

T7 for Medicare and Medicaid 

by: Sidney Trieger 
Trudi W. Galblum 
Gerald Riley 

Published by 

Department of Health and Human Sen/ices 

Health Care Financing Administration 

Office of Research, Dennonstrations, 

and Statistics 



Executive Summary 1 

Introduction 2 

The Legislation and the Programs 3 

Medicare and HMOs 3 

Medicaid and HMOs 4 

Current Knowledge 5 

Private Sector Research 5 

Utilization and Costs 5 

Prevention 6 

Quality 6 

Enrollment, Satisfaction, and Disenroliment 6 

Summary 7 

Research on Medicaid, Medicare, and HMOs 7 

Medicaid 7 

Utilization and Costs 7 

Quality of Care 8 

Enrollment, Satisfaction, and Disenroliment 9 

A Digression About California 10 

Medicare 11 

Summary 12 

Reimbursement 13 

Medicare 13 

Current Procedures 13 

Determination of Prospective Adjusted Average Per Capita Cost 14 

Experimental Reimbursement Procedures 15 

Selection Issue 16 

Medicaid 17 

Current Procedures 17 

Prepaid Health Research, Evaluation, and Demonstration (PHRED) and the Evolution of Rate- 
setting in California 18 

Marketing 20 

Medicare 20 

Demonstration Marketing Procedures and Results 20 

Conversions and Disabled Beneficiaries 21 

Medicaid 21 

PHRED Membership Studies 22 

Massachusetts Case Management 23 

Multnomah County (Oregon) Project Health 24 

Quality of Care 25 

Current Procedures 25 

HMOs and Professional Standards Review Organizations (PSROs) 26 

Medicare and Medicaid 26 

The PHRED Quality Assurance Demonstration 26 

Program Policy Considerations 27 

Medicare 27 

Claims Processing 27 

Coverage of Non-Medicare Services 28 

High and Low Option 28 

Premium for Low Option 28 

Institutionalized Beneficiaries 28 

Renal Beneficiaries— Coinsurance and Deductible 28 

Renal Beneficiaries— Master Record Annotation 28 

Cost Reporting 29 

HMO Medicare Manual 29 

Medicare Cards 29 

Medicaid 29 




Knowledge Gaps 29 

Medicare 29 

Medicaid 30 

Medicare and Medicaid 30 

Future Research 30 

Appendices 31 

A. Medicare Capitation Dennonstration Project Descriptions 

Fallon Community Health Plan 31 

Marshfield Medical Foundation 31 

Kaiser-Portland 31 

InterStudy Medicare Multiple Choice Program 32 

Health Central 32 

Kitsap Physicians Service 33 

Blue Shield of Massachusetts 33 

B. Institutionalization Data for Determination of the Adjusted Average Per Capita Cost 33 

Executive Summary 

The Health Maintenance Organization (HMO) Act of 1973 was aimed at making HMOs a viable alternative to fee- 
for-service care by mandating that, where feasible, employers of 25 or more individuals offer a choice between 
their traditional health insurance packages and at least one HMO. In recent years, much study has been devoted 
to understanding the impact of HMOs on utilization, cost, quality of care, and competition in the health care 
marketplace. Much of this research predates the HMO Act and focuses on under-65 employed persons enrolled 
in prepaid group practices under group contracts. 

Although Congress and the Department of Health and Human Services (DHHS) have generally supported the 
enrollment of Federal beneficiaries in HMOs, no more than 2 to 3 percent of these beneficiaries are HMO 
members. However, policymakers are becoming increasingly aware of HMOs' potential to contain health care 
costs without compromising quality of care. As means are sought to stimulate competition in the health care 
sector, HMOs are likely to become an important factor in whatever competitive approach is adopted. 
Consequently, there is renewed interest in promoting the expansion of Federal Hf^O contracting and Medicare 
and Medicaid enrollment. 

This paper provides a comprehensive overview of the major issues and alternatives which need to be addressed if 
public payer participation in HMOs is to expand. It focuses on current legislative barriers to HMO enrollment, 
knowledge about HMO benefits and liabilities, reimbursement problems and techniques, marketing strategies, 
quality assurance, administrative issues, and a future research agenda. 

Briefly, the paper supports the following conclusions with regard to HMOs and the Medicare program: 

• Medicare beneficiaries will enroll in HMOs if benefit packages are attractive. Over 20,000 beneficiaries have 
enrolled at only three demonstration sites. Almost all have enrolled through open enrollment rather than 
conversion. This represents 36 percent of all enrollees in HMOs from Medicare's 38 cost contracts. 

• HMOs will contract with the Health Care Financing Administration (HCFA) to enroll Medicare beneficiaries 
more readily if current risk and cost reimbursement procedures are modified. 

• Most HMOs should be able to provide a full benefit package at 95 percent of the adjusted average per capita 

• More flexibility in current HMO reporting requirements and elimination of certain Medicare coverage restric- 
tions (for example, spell of illness and physician extender use) would encourage more HMOs to contract with the 
Medicare program. 

• The open enrollment procedure needs to be assessed to determine if a health status adjustment is required 
for adjusted average per capita cost determination. The issue of whether HMOs favorably or adversely select 
Medicare enrollees remains a key question which is being addressed by HCFA's current demonstration 

The Medicaid demonstration projects discussed in this paper have yielded findings which should be valuable in 
encouraging State contracts with HMOs. These findings include the following: 

• Alternatives to door-to-door marketing can be effective in generating sufficient enrollments at a reasonable 
cost for most HMOs that would contract with a State. 

• States can implement ambulatory quality assurance systems which will provide the data necessary to monitor 

• States can use actuarial rate-setting methodologies to set rates for all categories of Medicaid recipients. 

• In a competitive marketplace, medically needy enrollees not otherwise eligible for Federal assistance will be 
willing to pay increased premiums to join the HMO of their choice; they will not necessarily select the lowest 
cost HMO available. 


On February 18, 1971, President Nixon issued a health 
message to Congress. In it, he advocated prepaid 
health care as a new approach to health services 
delivery in which the Federal government should be 
viewed as one of several partners. The message under- 
lined that the nation's health investment had reached 
nearly 8 percent of the GNP and announced a goal of 
1,700 HMOs with 40 million members by 1976. By June 
1976, only 1 percent of that goal (17 HMOs) had been 
realized. As of July 1980, 112 HMOs had been qualified 
by DHHS. A 10 year growth strategy developed by the 
Office of Health Maintenance Organizations now pro- 
jects 442 HMOs serving over 19.1 million members to 
be qualified by 1988.' 

Clearly, the HMO program has not yet fulfilled the 
hopes of its progenitors, and yet, many today believe 
that prepayment for health care might still be the most 
rational approach to restructuring the nation's health 
care delivery system. HMOs provide comprehensive 
health care reimbursed on a fixed per capita basis 
rather than fee-for-service. As a result of this payment 
mechanism, and an emphasis on health rather than the 
provision of services, HMOs have demonstrated a num- 
ber of advantages, such as: 

• a proven ability to reduce hospital admissions 

• the ability to provide accessible, comprehensive 
services and continuity of care 

• reduced out-of-pocket payments for health care 

• incentives to promote health maintenance and pre- 
vention as opposed to more expensive acute care 

• simplified claims procedures for both providers 
and beneficiaries 

• the ability to perform the above without a reduc- 
tion in quality of care 

• introduction of a competitive framework for the 
health care sector. 

Despite these benefits, success in attracting HMOs to 
contract with Medicare and Medicaid and in enrolling 
Federal beneficiaries in HMOs has been poorer than 
HMO development and penetration in the private 
sector. Supporters of HMO involvement with Medicare 
and Medicaid populations have had to face some 
special obstacles, ranging from legislative barriers to 
capitation reimbursement by the Medicare program to 
rapid turnover by Medicaid enrollees due to loss of 
eligibility. Except for a few demonstrations initiated to 
improve enrollment of Federal beneficiaries, even today 
no HMO delivers services to Medicare beneficiaries 
under reimbursement arrangements modeled after tra- 
ditional HMO management principles. However, the 
experience of these experiments in marketing to 
Medicare beneficiaries does strongly suggest that 
incentives such as increased benefits and reduced co- 
payments are effective in attracting the over-65 popula- 
tion at large to enroll. 

Unlike Medicare, Medicaid has not faced legislative 
obstacles to non-cost-based HMO reimbursement. In 
fact, just the opposite is true: current regulations pro- 
hibit Medicaid agencies from entering cost contracts 
with qualified HMOs. But the difficulties are consider- 
able in designing incentives to enroll categorical 
Medicaid recipients who already bear no out-of-pocket 
costs for medical care. In addition, many HMOs are not 
anxious to contract with States or serve the poor. 

This paper reviews the interaction of the Medicare and 
Medicaid programs with HMOs, with specific emphasis 
on what has been learned from research and demon- 
stration projects. Since much of the HCFA research is 
recent, particularly the Medicare demonstrations, the 
report should be considered as a snapshot look at the 
current state of the art. Preliminary findings from the 
implementation phases of these demonstrations, 
although necessarily tentative, have significance for 
current policy issues. 

The first section discusses the current legislative and 
reimbursement provisions for HMOs wishing to con- 
tract with HCFA to serve Medicare or Medicaid 
beneficiaries. The numbers of contracts and benefici- 
aries served are reviewed and the impediments to 
growth outlined. 

The second section contains an overview of the current 
knowledge about private sector and Medicaid experi- 
ence in HMOs, focusing on utilization and costs, 
prevention, quality of care, enrollment, satisfaction, 
and disenrollment. Because few articles have been 
published on Medicare and HMOs, each is reviewed. 

Section three provides a comprehensive review of 
Medicare HMO reimbursement issues, with emphasis 
on experience during the design phase of HCFA's 
seven Medicare risk-sharing demonstration projects. 
After outlining Medicare's current reimbursement pro- 
cedures, the section reviews the problems in setting 
prospective rates and discusses the various experi- 
mental methodologies. Special attention is paid to the 
adverse/favorable selection issue. It is this issue, 
perhaps more than any other, which has for the past 
several years influenced legislative support for 
Medicare participation in HMOs. 

Section four focuses on marketing and again is heavily 
weighted with experience from the Medicare demon- 
stration projects. It also reviews results from Cali- 
fornia's recent demonstration of alternative methods of 
enrolling Medicaid recipients and two other Medicaid 

Section five examines issues of quality assurance for 
Federal beneficiaries in HMOs. After reviewing current 
procedures and the status of HMOs' relationships with 
Professional Standards Review Organizations (PSROs), 
the section describes the quality assurance demonstra- 
tion and evaluation in California which ended in winter 
of 1981. 

"National HMO Development Strategy, 1978-1988,' 
USDHEW, OHMO. Rockville, Md. 

The sixth section reviews the various administrative 
issues which arose during the design of the Medicare 
demonstrations. This experience should be useful in 
refining HCFA's systems and procedures for managing 
Medicare HMO contracts. 

Finally, the authors recommend directions for future 
research on Federal beneficiary enrollment in HMOs. 

The Legislation and tlie Programs 

The HMO act of 1973 substantially increased the ability 
of HMOs to compete with traditional health insurers by 
requiring that all employers covered by the Fair Labor 
Standards Act of 1938 and employing 25 or more offer 
at least one "group" and one "independent practice 
association" HMO as an alternative to whatever other 
health insurance package it offers to employees.^ The 
employer must cover up to what it would pay for a 
non-HMO health insurance subscriber. In order to 
benefit from this "dual choice" provision of the law, 
HMOs must seek and be granted Federal qualification. 
This means, among other things, that they must offer a 
minimum benefit package provided under a community 
rating, and they must have at least one 30-day period 
per year during which any individual can apply and be 
accepted for membership. 

Although some believe that the requirements for 
Federal qualification are unduly restrictive, the belief 
that government has a special responsibility to protect 
Federal beneficiaries from unscrupulous health care 
providers has led to minimal flexibility with regard to 
reimbursement, benefit packages, and structural 
characteristics required for an HMO wishing to con- 
tract to serve Medicare and/or Medicaid beneficiaries. 
This view was reinforced in 1975, after the unethical 
behavior of some prepaid health plans (PHPs) with 
Medicaid contracts in California was revealed in Con- 
gressional hearings. While the HMO Amendments of 
1976 eliminated some of the worst barriers to HMO 
qualification, they signaled a strong orientation toward 
Medicare and Medicaid contracts with qualified HMOs 
only. For States like California, which at one point had 
53 contracts with PHPs to serve over 250,000 Medicaid 
recipients, at a cost of $84.6 million per year, this 
meant the demise of all but 13 contracts. 

Initiatives to revitalize the HMO movement in 1977 
under Secretary Joseph Califano and Under-Secretary 
Hale Champion alleviated some of the factors which 
had slowed the development of HMOs, including lax 
Departmental management, administrative backlog in 
qualifying new HMOs, and excessive delays in issuing 
regulations. The HMO Amendments of 1978, enacted in 
November, resulted in a three year extension of the 

2. A "medical group" is composed of a closed panel of physi- 
cians who pool their income and share common facilities, 
support staff, and medical records. A fee-for-service group 
that wishes to qualify as a "medical group" must, as its 
principal professional activity and group responsibility, 
engage in the coordinated practice of the profession for an 
HMO. The HMO Amendments of 1976 define this to mean 
that at least 35 percent of the group's patients must be pre- 
paid. An "independent practice association" is composed 
of an open panel of fee-for-service solo practitioners in a 
group which agrees to provide prepaid care to plan 

HMO program authorizations, increased and expanded 
support for developing HMOs in the form of grants, 
contracts, and loan guarantees, and introduced an 
HMO management intern program. However, the 
provisions of the bill introduced by Senator Richard 
Schweiker in February 1978 to stimulate meaningful 
participation of the poor and elderly in HMOs were 

Hale Champion declared in a June 1977 policy speech 
at the Group Health Association of America conven- 
tion: "Secretary Califano doesn't think HMOs are 
demonstration projects. He thinks they are for real." 
They may be for real, but for Medicare and Medicaid in 
1980, they remained a footnote. 

Medicare and HMOs 

Prior to the enactment of the 1972 Amendments to the 
Social Security Act, Medicare beneficiaries enrolled in 
HMOs could receive prepaid services covered only 
under Part B from Group Practice Prepayment Plans 
(GPPPs). GPPPs are reimbursed on the basis of their 
reasonable costs. In March 1970, the Administration 
proposal known as the "Part C" fixed price prepayment 
option was added by the House Ways and Means 
Committee to a marked up welfare reform bill. At the 
same time, HEW issued a press release which included 
the following statement: 

Our goal, the Secretary said, is that every elderly or 
poor person, covered by Medicare or Medicaid, be 
given the right to choose between receiving services 
under such a contract and receiving individual hospi- 
tal and physician services in the traditional manner. 
We must promote diversity, choice, and health com- 
petition in American Medicine if we are to escape 
from the grip of spiralling costs.^ 

However, the proposal, as ultimately enacted in 
Section 1876 of the Social Security Act (Section 226 of 
P.L. 92-603), little resembled the House version (H.R. 
1). Viewing the proposal as an invitation to HMOs to 
reap excessive savings from the Medicare trust funds, 
the Senate Finance Committee prevailed in retaining in 
the legislation the essential elements of cost reim- 
bursement. Final regulations implementing Section 
1876 were not published until late 1976. 

Section 1876 provides for two methods of reimburse- 
ment to HMOs, both involving a single capitation rate 
for services covered under Parts A and B of Medicare. 
A mature HMO may enter into either a cost or a risk basis 
contract." Cost contracts permit reimbursement only 

3. Press release, March 25, 1970. 

4. A mature HMO is defined in 45 CFR 405.2001(b)(1) as one 
which enrolls at least 5,000 prepaid members, provides all 
Part A and B services available to non-members in the ser- 
vice area, and meets all applicable Federal qualifying 

To enter into a risl< contract, a mature HMO must enroll at 
least 25,000 prepaid members, if it serves an urban area, 
and at least 5,000 if it serves a non-urban area (45 CFR 
2004(a)(1) and (a)(2). A developing HMO may enter into a 
cost contract only, and must demonstrate to the Secretary 
that it is committed to meeting all of the requirements for a 
mature HMO within three years after the effective date of 
its initial contract (45 CFR 405.2001(a)(2)). 

for the reasonable costs of providing covered items or 
services to Medicare enrollees. Reimbursement under a 
risk contract is based on a comparison of the HMOs' 
reasonable incurred costs and the Adjusted Average 
Per Capita Cost (AAPCC). The AAPCC is the average 
cost of providing covered items and services under fee- 
for-service in the same area as the HMO for the same 
demographic composition of Medicare beneficiaries 
enrolled in the HMO. (See the section on reimburse- 
ment for a more detailed discussion.) Under risk 
reimbursement, the Medicare beneficiary is "locked-in" 
to the HMO so that s/he may only receive services 
approved by the HMO. An HMO providing services 
under a risk contract is financially liable for covered 
services and items furnished to its Medicare enrollees 
outside the HMO, even in the absence of the HMO's 
prior approval, where such services are emergency or 
urgently needed, or are Medicare covered and deter- 
mined by the Secretary not to have been made 
reasonably available by the HMO to its Medicare 

As of July 1980, the Medicare program had 38 cost 
contracts and one risk contract signed with HMOs to 
serve approximately 61,500 enrollees (representing less 
than 1 percent of the total Medicare population). 
Clearly, the lack of HMO penetration into the over-65 
population can be traced, in part, to the above reim- 
bursement provisions. If an HMO chooses cost 
reimbursement, it can broaden its sources of revenue 
at the price of filling out Federal cost reports and com- 
plying with other requirements for Federal qualification. 
This grafting of cost reimbursement on to HMOs, how- 
ever, forces an awkward compromise on the HMO 
geared to operate under a prospective budget. The 
HMO that chooses a risk contract also must fill out the 
cost reports and be Federally qualified. But in addition, 
it runs the risk of adverse selection, as well as the 
responsibility for absorbing losses resulting from its 
own inefficiencies. 

In addition to the impediments to development of 
HMOs in current reimbursement policies, marketing 
the HMO to the Medicare beneficiary is a problem. 
From the HMO's perspective, it is costly because Medi- 
care beneficiaries are not part of a group and are diffi- 
cult to reach. From the beneficiary's perspective, there 
are disincentives to enroll, such as the "lock-in." HMO 
enrollment for the Medicare beneficiary also means 
incurring the same deductibles and coinsurance costs 
for no richer array of benefits than they are entitled to 
under Parts A and B of Medicare. 

Medicaid and HMOs 

The HMO was initially developed to deliver health care 
to employed persons and their families. As enacted in 
1965, Title XIX did not provide for the enrollment of 
Medicaid eligibles in HMOs. The first efforts to enroll 
low income persons in HMOs were supported by the 
Office of Economic Opportunity (OEO). HMOs such as 
the Health Insurance Plan (HIP) of New York, Group 
Health Cooperative (GHC) of Puget Sound, and 
various Kaiser plans were awarded Neighborhood 
Health Center grant funds to enroll low income 

persons not necessarily eligible for Medicaid. The 
grant funds also provided services such as child care 
and social services not normally covered under Title 
XIX. Thus both the HMOs and low income populations 
had incentives to participate through guaranteed eligi- 
bility and enriched benefit packages. 

The legislative authority for States to contract with 
HMOs is contained in Section 1902(a)(23) of the 1967 
Amendments to the Social Security Act (Section 227(a) 
(3)). Known as the "freedom of choice" provision, it 
recognizes an "organization which provides services, 
or arranges for their availability on a prepayment 
basis." Regulations for implementing this legislation 
provided little guidance other than (1) that the State 
specify the amount of the premium, the services 
covered, and the term of the contract; (2) that the 
premium payment fully discharges the State from 
responsibility for the costs of the covered services; (3) 
that the premium amount and/or covered services be 
periodically renegotiated; and (4) that the State require 
the prepaid health plan to maintain and provide such 
records and reports necessary for the State to meet 
Federal reporting requirements. The State, in order to 
obtain Federal sharing, had to pay the plan premiums 
which would not exceed the cost of providing the same 
services under the fee-for-service system. Prior Depart- 
mental approval of HMO contracts was not required. 
The 1972 Amendments to the Act amended Section 
1902(a)(23) to allow States to contract with an organi- 
zation to provide care and services in addition to those 
under the State plan in a sub-state area, thereby 
avoiding conflict with the state-wide and comparability 
provisions of the law. 

Twelve States, including the District of Columbia, 
signed 66 contracts with HMOs between 1971 and 1973 
to serve a total enrolled population of about 371,000 
Medicaid eligibles.' Fifty of these contracts were imple- 
mented in California under the Medi-Cal Reform Act of 
1971. But within the first year of this program in 
California, serious complaints were raised about the 
cost, quality, enrollment practices, and corporate 
accountability of the PHPs contracting with the State. 
Failure by both the State and the Department to ade- 
quately address these complaints culminated in 
hearings in 1975 held by the Senate Permanent Sub- 
committee on Investigations of the Committee on 
Governmental Affairs. One of the chief charges 
directed at HEW in those hearings was that the 
Department had yet to issue regulations specifying the 
procedures for obtaining a determination of whether an 
HMO should be "qualified" under the HMO Act of 
1973. The regulations were finally published on August 9, 
1975, placing a number of new requirements on State 
contracts with HMOs. These included (1) that the 
State control HMO enrollment and disenrollment prac- 
tices; (2) that services be available on a 24 hour, seven 
days a week emergency basis; (3) that internal 
grievance procedures be established; and (4) that a 
medical record-keeping system exist. The regulations 

5. State Medicaid HMO Contracts, USDHEW, Social and 
Rehabilitation Service, Medical Services Administration 
(SRS-74-24807), September 1973. 

also expanded access to records requirements by 
stipulating that the State and the Department have the 
right to inspect and evaluate the quality, appropriate- 
ness, and timeliness of services provided and audit and 
inspect any books and records of the HMO v^rhich 
relate to the contract. The State also had to provide 
HEW with the actuarial basis for the premium deter- 
mination, and the Department required prior approval 
of contracts with expected values of $100,000 or more. 

The HMO Amendments of 1976 went further toward 
correcting the conditions which permitted the scandals 
in California to occur. Title XIX matching funds were 
limited to Federally qualified HMOs with few excep- 
tions, and qualified HMOs were restricted from 
enrolling more than 50 percent of their members from 
Medicare and Medicaid. Although the 50 percent 
requirement may ensure quality service provision for 
Federal beneficiaries, it also establishes a new access 
barrier to HMOs for beneficiaries residing in inner city 

Utilization and Costs 

From his exhaustive assessment of the available litera- 
ture on HMOs, Harold Luft concluded that members of 
staff model HMOs/prepaid group practices (PGPs) 
have the lowest costs, as compared to members of 
independent practice associations (IPAs) and major 
medical-indemnity plans.' Luft determined that annual 
costs for Blue Cross/Blue Shield subscribers were 16 
to 88 percent higher than for enrollees in the lowest 
cost PGPs. In addition, average out-of-pocket costs per 
person and per family for HMO enrollees were less, 
particularly in staff as opposed to IPA models. We do 
not know, however, whether these cost differences 
represent true cost savings because of the self- 
selection factor. That is, those who choose to enroll in 
an HMO may be different from the general population 
on one or more parameters which have a proven corre- 
lation to the use of health care. HMO enrollees may be 
healthier, or they may seek less health care, regardless 
of their health conditions. 

As of 1980, 53 HMOs in 17 States had contracts to 
serve Medicaid recipients.^ Twelve of these contracts 
are in California, seven in Maryland, five in Michigan, 
and four each in Massachusetts, New York, and 
Minnesota. Approximately 270,000 Medicaid recipients 
are enrolled in HMOs, representing about 2 percent of 
total eligibles and 6.7 percent of the total enrollment of 
the plans. Although contracts with qualified HMOs 
increased from 24 to 27 between 1978 and 1980, Medi- 
caid enrollment has dropped nearly 15 percent. A 
strong potential for further reductions in enrollments 
exists due to the large number of current contracts 
with non-qualified plans. Only 55 percent of current 
contracts are with qualified HMOs; the 1976 amend- 
ments permit States to contract with non-qualified 
HMOs under limited defined circumstances or for a 
three year period when the plan is seeking Federal 
qualification. Should many of the 24 non-qualified con- 
tracting plans not receive qualification, their contracts 
would be terminated. 

Current Knowledge 

Private Sector Research 

The paucity of Medicare and Medicaid beneficiaries 
enrolled in HMOs has yielded a commensurate dearth 
of research on cost, utilization, marketing, and quality 
of care delivered to the poor and elderly in HMOs. To 
date, most of the research on HMO performance has 
been based on private sector data, particularly from a 
small number of HMOs which have consistently partici- 
pated in research studies. This research has produced 
some generally accepted conclusions about HMOs, as 
well as some clear indications about where more study 
is needed. The following section consists of a brief 
review of the state of the art, focusing on those aspects 
of research which are relevant to Medicare and 
Medicaid involvement in HMOs. 

The chief means by which HMOs control costs is 
reduced hospitalization. Luft found that in 44 of the 57 
comparisons of HMO to fee-for-service experience, 
dating back to 1951, HMO enrollees had fewer hospital 
days than the comparison group. In 46 cases, the 
admission rate was lower. HMOs do not appear to have 
a significant effect on length of stay (LOS). Out of the 
57 cases, 30 showed lower LOS, six were the same, 
and 21 were higher. Case-mix adjustments do not alter 
this finding. Overall, the utilization of staff model 
HMOs is about 35 percent less than comparison 
groups, while IPAs are about 5 to 25 percent lower. As 
to whether these results are associated with self- 
selection, the evidence is mixed. Some studies indicate 
that those likely to be high users tend to opt for con- 
ventional health insurance plans. On the other hand, 
some argue that persons who anticipate a high need 
for health care are more likely to choose an HMO. 

Studies about the relationships of HMOs and ambula- 
tory care are less conclusive than those about hospital- 
ization. The HMO rhetoric frequently refers to the 
substitution of less expensive outpatient care. But 
more study is needed to understand the dynamics of 
this: Do HMOs eliminate unnecessary care, do they 
underserve, do they selectively enroll, or do they sub- 
stitute ambulatory care? The only sure conclusion 
which Luft could draw from his analysis was that a 
larger proportion of HMO enrollees have at least one 
visit per year compared to non-enrollees. He also was 
reasonably confident that, while HMO enrollees have 
more ambulatory visits per year than people in com- 
parison groups, the difference is less than 10 percent, 
and nearly as many HMOs show fewer visits per year 
as show more. The actual extent of substitution of 
physician visits and ancillary services for inpatient care 
is not known. 

6. Most of the statistics quoted here were prepared by 

HCFA's Division of Alternative Reimbursement Systems in 
a report entitled "Medicaid Experience Summary with 
HMOs and Other Prepayment Plans." 

7. Luft, H. S., Health Maintenance Organizations: Dimensions 
of Performance (A Wiley-lnterscience Publication: John 
Wiley & Sons, New York), in publication. 


One of the chief benefits that HMOs claim is that they 
respond to incentives to provide preventive care 
because in doing so they save the costs they would 
have otherwise incurred for acute care. But studies 
comparing the use of preventive services by HMO 
enrollees and non-enrollees produce conflicting results. 
Contrary to popular expectation, non-enrollees used 
preventive care more or as much as HMO enrollees in 
four out of 11 studies analyzed by Luft. The explana- 
tion appears to be that demand for preventive services 
is more a function of coverage than provider 

Analysis of preventive care is beset by some thorny 
problems, such as defining preventive care and under- 
standing the role of the physician in his or her choices 
of services. Furthermore, intuitive assumptions that 
preventive care is always good are subject to challenge 
in terms of efficacy as well as economic costs and 
benefits. However, two perspectives are available from 
which to consider the provision of preventive services 
in HMOs. From the HMO's perspective, there are clear 
economic incentives to discourage the unnecessary 
provision of discretionary, preventive services; from the 
enrollee's perspective, the elimination or reduction of 
out-of-pocket costs for ambulatory care appears to act 
as an incentive to seek preventive services. The value 
and appeal of preventive care to the elderly has 
received little, if any, research attention to date. 
Assuming prevention is valuable, we have very little 
information about current utilization of such services 
by Medicare beneficiaries or how more comprehensive 
coverage would affect utilization. 


Two comprehensive reviews of the literature on quality 
of care delivered by HMOs have been performed.' ^ 
Both recognize that the state of the art limits the cer- 
tainty of conclusions, and both exonerate HMOs in 
general from the allegation that they underserve 
enrollees to achieve economies. 

The first review, by Luft, is organized on the basis of 
structure, process, and outcome studies. With regard 
to structure, Luft concludes that HMOs are at least as 
good as fee-for-service: HMOs seem more likely to re- 
cruit and attract more certified specialists (although 
the superiority of such credentials is unproven), admit 
to accredited hospitals, and provide more continuing 
education to their staff. On the other hand, arguments 
that physicians in HMOs more frequently consult with 
each other was not supported by Luffs review. Luft 
also found that while internal peer review is present in 
most HMOs, it is not found in all of them. Where infor- 
mation on quality is available, it is not clear whether it 
is used or is effective in instituting improvements. 

The relationship of the process of care to quality is 
nearly as tenuous as the relationship of structure to 
quality. Luft makes the point that assessments of 
quality delivered in HMOs based on process measures 
are easily biased in favor of settings which keep good 
records and offer an array of technical services. Thus, 
HMOs appear to do better on process measures which 
pick up lab tests and procedures, but this could be 
more due to coverage than quality differences. Studies 
based on HMO outcomes are quite few and of limited 
value due to small sample sizes. However, the prepon- 
derance of what is available suggests that HMO out- 
comes are not significantly different than fee-for- 

The conclusions of the second review, by Frances C. 
Cunningham and John W. Williamson, are more posi- 
tive than Luffs. The authors analyzed 25 studies in 
which they identified 34 different measures of quality 
(seven outcome, 25 process, and two structure). The 
studies reported a total of 84 quality measurements, of 
which 65 were considered valid for this review. Of the 
65 measures, care provided by HMOs appeared 
superior in 50 cases, similar in 14, and inferior in one (a 
Medicaid population). The authors concluded that the 
quality of care provided by HMOs is comparable, if not 
superior, to conventional settings. 

Enrollment, Satisfaction, and Dlsenrollment 

The literature on satisfaction reviewed by Luft indicates 
that HMO enrollees are more satisfied with their finan- 
cial coverage than non-enrollees. Thus, while a person 
who chooses HMO enrollment because of its better 
financial coverage may subsequently disenroll because 
of dissatisfaction with something else, financial cover- 
age remains a key means to motivate enrollment. The 
significance of this for Medicare and Medicaid is that 
these programs can do little within their current 
authority to motivate enrollment via financial 

Aside from economic incentives, studies about why 
people choose HMOs focus on their feelings about the 
care they receive, out-of-plan utilization, and dlsenroll- 
ment. Luft calls the latter two "behavioral correlates of 
satisfaction." Overall, Luft found that out-of-plan utili- 
zation accounts for 7 to 14 percent of all services 
received by HMO enrollees. Outside users are the ones 
who most frequently express dissatisfaction in surveys. 
The percentage of those who disenroll annually is 
usually under 10 percent. Curiously, some plans with 
the lowest dlsenrollment rates do more poorly in 
measures of consumer satisfaction than plans with 
less stable enrollment. Hirschman has shown that 
HMO enrollees are generally more informed 
consumers, and as such, they may be more vocal in 
their complaints.'" Nevertheless, men the complainers 
usually do not disenroll, probably because they enjoy 
the coverage they receive at a reasonable premium. 

8. See Footnote 7. 

9. Cunningham, F. C. and J. W. Williamson, "How Does the 
Quality of Health Care in HMOs Compare to That in Other 
Settings? An Analytic Literature Review: 1955 to 1979." 
The Group Health Journal, Winter 1980, pp. 4-13. 

10. Hirshman, A. O., Exit, Voice and Loyalty: Responses to 
Decline in Firms, Organizations, and States, (Cambridge: 
Harvard University Press), 1970. 

Since the Medicaid and Medicare programs cannot 
offer financial savings to encourage beneficiaries to 
join HMOs, it is important to consider what other fac- 
tors might motivate enrollment. These factors might 
include certain benefits which HMO members presum- 
ably enjoy, such as better accessibility. Luft found that 
while HMOs offer shorter office waiting times, waiting 
times for appointments are longer. For elderly people 
with urgent needs who visit the doctor frequently, this 
could be a significant deterrent to HMO membership. 
For the low income person, this may be less important 
than simply having a health provider in the vicinity who 
accepts Medicaid patients. 

Continuity of care is often assumed to be more readily 
available to HMO members. The empirical benefits of 
continuity of care are by no means clearly defined; 
however, there is some consensus that continuity is a 
necessary component of quality. Without the availability 
of a longitudinal medical record and/or a physician 
who knows the patient, it is presumed that effective 
prevention, identification, and treatment of disease 
cannot be achieved. It appears that HMOs provide less 
opportunity for members to identify with a personal 
physician but possibly better maintenance of medical 
records.^' Self-selection may partially account for this 
if HMOs automatically attract people who have and 
seek no personal physician relationship. By the same 
token, however, many elderly persons highly value this 
relationship. Unless the HMO offers an appealing 
substitute to their current doctor, they may not enroll. 
Furthermore, since the evidence shows that HMO 
enrollees are less satisfied with doctor-patient commu- 
nication and relationships, disenrollment among the 
elderly may prove higher. 


With minimal equivocation, researchers attribute HMOs' 
lesser costs to reductions in hospitalization. To what 
reduced hospitalization should be attributed, however, 
is part of an important controversy about self-selection 
which is discussed elsewhere in this report. The effect 
of HMOs on the use of ambulatory and preventive care 
has less consensus than HMOs' effect on inpatient use, 
due in part to the wide variety of services encom- 
passed by ambulatory preventive care. One can find 
studies which argue that HMOs provide more or less of 
such services. But the theory that HMOs provide more 
to prevent future illness is treated with increasing skep- 
ticism. To the extent that quality of care lends itself to 
measurement, no consistent evidence of lesser quality 
in HMOs has yet been produced, while there is some 
research suggesting that HMOs may offer improved 
quality over fee-for-service medicine. Studies about 
consumer attitudes on health insurance show that 
HMO enrollment affords more satisfaction with 
financial coverage than do fee-for-service plans. Other 
factors which may play a role in HMOs attracting and 
retaining members are accessibility and continuity of 

Research on Medicaid, IVIedicare, and Hiy/IOs 


Medicaid and related programs have had a broader 
experience with prepaid contracting than the Medicare 
program. This has resulted in three types of studies on 
the subject: (1) studies which deal with low income 
(such as Office of Economic Opportunity (OEO)) 
populations not necessarily on public assistance, (2) 
studies concerned with Medicaid eligibles, and (3) 
analyses of the California Medi-Cal experience. For a 
complete review of the literature about Medicaid as 
well as low income persons in HMOs, Margo L. Vignola 
and George B. Strumpf have prepared an annotated 
bibliography entitled: Medicaid Eligibles in Health 
Maintenance Organizations: Utilization, Cost, Quality, 
Legal Requirements (December 1977, updated in 1980). 
Below is a brief overview of the major issues regarding 
Medicaid and HMOs. 


Since the inception of Medicaid, a benchmark of the 
program's success has been how well it has brought 
the poor into the mainstream of health care. This goal 
assumes that patient needs are similar, regardless of 
socio-economic status. While a number of studies 
compare OEO and other low income populations' utili- 
zation of HMOs to middle class enrollees in the same 
HMO, only one such study is based on Medicaid 
recipients. ^^ This study compared the utilization experi- 
ence of Medicaid and non-Medicaid low income 
enrollees in the Harvard Community Health Plan 
(HCHP) to employer group enrollees. The researchers 
found that Medicaid enrollees cost less and used fewer 
hospital days per 1,000 than group enrollees (234.5 
versus 286.4), while low income days per 1,000 were 
significantly higher (462.1) than days for Medicaid or 
group enrollees. Medicaid and group utilization and 
costs for ambulatory care were comparable, but low 
income populations' ambulatory utilization, again, was 
higher, with the exception of adult females. 

There are two well-known studies which compare 
Medicaid enrollee utilization in HMOs to Medicaid fee- 
for-service. One is based on the experience of Group 
Health Association (GHA) in Washington, D.C., and the 
other on an analysis of 10 HMOs (six in California, 
Group Health Cooperative of Puget Sound in 
Washington, HCHP in Massachusetts, HIP of Greater 
New York, and the Temple Health Plan of 
Pennsylvania).^^ ^^ 

11. Richardson, W. C, S. M. Shorten, P. K. Diehr, "Access to 
Care and Patient Satisfaction," In William C. Richardson, 
(editor), The Seattle Prepaid Health Care Project: Com- 
parison of Health Services Delivery, Seattle: University of 
Washington, School of Public Health and Community 
Medicine, 1976. 

12. Coltin, C. C, R. Neisuler, and R. S. Luria, Evaluation of a 
Program to Facilitate the Integration of a Low Income 
Population Into a Prepaid Group Practice, (Harvard Com- 
munity Health Plan: Cambridge, July 1976). 

13. (1) Fuller, N. and M. Patera, Report on a Study of Medi- 
caid Utilization in a Prepaid Group Practice Plan, 
(DHEW/PHS/MSA) Washington, D.C., January 1976; and 
(2) Fuller, N., M. Patera, and K. Koziol, "Medicaid Utiliza- 
tion of Services in a Prepaid Group Practice Health Plan," 
Medical Care, September 1977, pp. 705-737. 

14. Gaus, C, B. Cooper, and C. Hirschman, "Contrasts in 
HMO and Fee-For-Service Performance," Social Security 
Bulletin, May 1976, pp. 3-14. 

The GHA study is based on 1,000 Medicaid recipients 
enrolled between 1972 and 1974 and a control group 
composed of the AFDC and disabled Medicaid universe 
in Washington, D.C. for fiscal year 1972. It excluded 
individuals under age 1 and over age 65. Both inpatient 
and ambulatory utilization were lower after enrollment 
and lower than that of the control group. Despite inclu- 
sion of four additional benefits (dental, outreach, pros- 
thetic, and ambulance), the costs per enrollee did not 
increase over three years and actually decreased 
between 1972 and 1973. This contrasts with a control 
group annual cost increase per person from $373 in 
1972 to $465 in 1974. 

Unlike some HMO studies, the GHA study did look at 
the incidence of out-of-plan utilization. Claims that 
HMOs reduce utilization and costs cannot be fully sub- 
stantiated without knowing the extent to which enroll- 
ees have used out-of-plan services, whether they were 
paid for, and by what party. The GHA study found out- 
of-plan use to be quite low, although there were signi- 
ficant out-of-plan purchases of prescription drugs. 
Barring abuses, these services were paid for out-of- 
pocket; Medicaid recipients who joined GHA sur- 
rendered their Medicaid cards. 

The other widely known study about Medicaid utiliza- 
tion in HMOs compared the experience of 8,000 Medi- 
caid families in 10 HMOs to fee-for-service utilization. 
Using a matched control group, the study found that 
hospital utilization and the incidence of surgery were 
less than fee-for-service for prepaid group practices 
but that there were no significant differences in utiliza- 
tion between foundations for medical care (that is, 
IPAs) and fee-for-service. The authors concluded that 
organized groups of salaried physicians may exert 
more influence on utilization than capitation payments 
alone. This study did not assess cost differences. In 
ambulatory care, the study showed no differences 
between HMO and fee-for-service controls. The 
authors posited this, as did the GHA study, as evidence 
that reductions in hospitalization are not achieved by 
substituting ambulatory care. 

This study also reported minimal out-of-plan usage- 
less than 1 percent. Other research is showing that a 
significant proportion of Medicaid enrollees do use 
out-of-plan services. A study of 1972 utilization data for 
HIP Medicaid enrollees showed that the Medicaid 
enrollees were seeking one-half to two-thirds fewer 
physician visits than regular HIP subscribers.'^ How- 
ever, after analyzing a sample of claims submitted by 
fee-for-service providers to see how many services 
they delivered to HIP Medicaid enrollees, it turned out 
that two-thirds of the physician services the enrollees 
were receiving were delivered out-of-plan. The study 
did not report how many of these claims were actually 
paid by the State. Further, since the HIP contract did 
not include all of the services covered by the State, not 
all of the out-of-plan utilization was inappropriate. 

A recently completed analysis of out-of-plan utilization 
by 425 AFDC recipients enrolled in "a large HMO in 
Massachusetts" between April 1977 and June 1978, 
however, suggests that HIP's experience may not be 
unique.'^ Thirty-eight percent of the AFDC enrollees in 
this HMO used non-emergency, non-urgent, out-of- 
plan ambulatory services during a one year period, 
with women more likely to do so than their children. 
The length of pre-enrollment eligibility was unrelated 
to out-of-plan utilization during enrollment. In addition, 
the types and patterns of use of out-of-plan providers 
were similar before and after enrollment. However, 
high users of ambulatory care before enrollment did 
appear more likely to use out-of-plan services after enroll- 
ment than low users before enrollment. The authors 
suggest that HMO marketing and enrollment of 
Medicaid recipients who have established relationships 
with fee-for-service providers may not be appropriate 
and that better education of enrollees on proper use of 
the HMO may be helpful. 


With quality of care measurement at its present level of 
sophistication, the extent of preventive care provided is 
frequently used as a proxy for quality. Studies which 
compare the use of preventive services between Medi- 
caid and fee-for-service populations produce contra- 
dictory results. Analyses of the East Baltimore Medical 
Plan indicate that higher proportions of enrolled 
children receive some preventive care than children 
treated by fee-for-service doctors and that enrolled 
children are more likely to receive care.'' '^ In the GHA 
study, in which Luft identified seven types of preventive 
services, use of these services was significantly higher 
for the enrolled group; however, the proportion of 
children under six who had ever received each of five 
immunizations and the percentages of women who 
received pre- and post-natal care were consistently 
lower for the enrolled populations. 

Measures of maternity care used in the study com- 
paring 10 HMOs to fee-for-service care also showed 
that 52 percent of enrolled women with live births had 
11 or more prenatal visits compared to 60 percent in 
the controls.'^ The authors concluded that HMOs do 
not provide more preventive care than fee-for-service 
care; at no site was preventive care greater in HMOs, 
and in some it was significantly less. 

15. Hester, J. and E. Sussman, "Medicaid Prepayment: Con- 
cept and Implementation," Milbank Memorial Fund 
Quarterly/Health and Society, Fall 1974, pp. 415-444. 

16. Nassif, D., S. Smith, D. Sisson, and C. Greenfield, "Out of 
Plan Utilization by Medicaid Recipients Enrolled in an 
HMO Program," presented at the 108th APHA Annual 
Meeting, October 1980. 

17. German, P., E. Skinner, S. Shapiro, and D. Salkever, 
"Preventive and Episodic Health Care of Inner-City 
Children," Journal of Community Health, pp. 92-106, 1976. 

18. Salkever, D. S., P. S. German, S. Shapiro, R. Morky, and 
E. A. Skinner, "Episodes of Illness and Access to Care in 
the Inner City: A Comparison of HMO and non-HMO 
Populations," Health Services Research 11, pp. 252-270, 

19. See Footnote 14. 

In addition to looking at preventive services, some 
studies attempt to draw conclusions about quality from 
mortality and utilization statistics. In 1967, Shapiro 
found that the proportion of indigent aged persons 
who received no physician services after enrolling in 
HIP in 1962 went down compared to no change in the 
non-HIP group. 2° But previous high users of care who 
enrolled averaged fewer physician visits in HIP than 
under fee-for-service. The death rate for ambulatory 
enrollees in the 18 months following the study year 
was 14 percent lower than the comparison group. 

Another way of addressing quality of care is to deter- 
mine whether services are accessible in the first place. 
Since a major HMO selling point for Medicaid recipi- 
ents is accessibility to providers who will always treat 
them, it is particularly important to know if HMOs are 
adequately delivering "accessibility." Although there 
are no studies on whether the same HMO is as acces- 
sible to its Medicaid enrollees as it is to its employed 
groups, there are a number of studies which compare 
the accessibility of care to Medicaid enrollees and fee- 
for-service Medicaid control groups. Waiting time in 
HMO offices is uniformly shorter than for fee-for- 
service doctors, but in some cases, it takes longer to 
get an appointment. Other measures of accessibility 
with conflicting results include telephone access, 
general convenience, time allocated with physician for 
appointment, and physical access. 

LoGerfo et al conducted a study under the Seattle 
Prepaid Health Project which concerned low income as 
opposed to categorically needy persons. The study 
found consistently higher compliance with optimal care 
criteria for urinary tract infections for enrollees in the 
Group Health Cooperative of Puget Sound compared 
to those receiving care from physician members of 
King County Medical/Blue Shield county-wide medical 
services.^^ Furthermore, higher surgical rates for four 
common procedures in the fee-for-service group were, 
in part, attributable to a higher proportion of cases 
which did not meet "common criteria" for surgery." 

On the neutral side, the study of 10 HMOs showed that 
HMO enrollees averaged 1.3 days of disability per 
month compared to controls which averaged 1.4 days, 
although the authors concede that no direct attempt 
was made to measure health status.^^ Another study 

20. Shapiro, S., J. J. Williams, S. Jerley, P. M. Densen, and H. 
Rosner, "Patterns of Medical Care by the Indigent Aged 
Under Two Systems of Medical Care," American Journal 
of Public Healtti 57, pp. 784-790, 1967. 

21. LoGerfo, J. P., J. Larson, and W. C. Richardson, "Assess- 
ing the Quality of Care for Urinary Tract Infection in 
Office Practice: A Comparative Organizational Study," 
Medical Care 16, pp. 488-495, 1978. 

22. LoGerfo, J. P., R. A. Efird, P. K. Diehr, and W. C. 
Richardson, "Rates of Surgical Care in Prepaid Group 
Practices and the Independent Setting: What are the 
Reasons for the Differences?" Medical Care 17, pp. 1-7, 

with inconclusive results was conducted by Louis and 
McCord.^" The study was undertaken amid the uproar 
in the mid-1970s concerning the entire management of 
the PHP program in California. The authors found the 
PHPs to be structurally adequate in most respects 
except peer review. They evaluated process of care to 
determine if basic health assessment and maintenance 
services were being provided and observed substantial 
variations between PHPs. Fee-for-service providers 
scored higher on quality of infant care, while the PHPs 
did better for school-aged children and adults. 
Analyses of disease specific processes again showed 
wide variations between PHPs, some better, some 
worse than fee-for-service care. The authors most 
pointed conclusion was that the quality of care pro- 
vided to Medi-Cal recipients was disappointingly low, 
regardless of whether it was fee-for-service or PHP. 


Marketing HMOs to Medicaid recipients is difficult due 
to a number of factors. Many recipients simply are not 
aware that they have a choice. If they are, choice may 
seem unimportant because of the more immediate non- 
health related problems that welfare applicants must 
cope with. Even if the choice is perceived as important, 
adequate information on specific HMOs— their location, 
their hours of operation, etc.— may not be readily avail- 
able. Furthermore, since Medicaid eligibility is not 
assured for a fixed length of time, the HMOs may not 
elect to market aggressively. 

If the Medicaid program is ever to enroll recipients in 
substantial numbers in HMOs, policymakers need a 
better understanding of the key factors which motivate 
this choice. In Thomas Bice's study of the East Balti- 
more Medical Plan (EBMP), dissatisfaction with fee-for- 
service care was related to Medicaid recipients' deci- 
sion to enroll." At the same time, expressions of dis- 
satisfaction with fee-for-service by enrollees were not 
intense. There are at least two possible explanations 
for this anomaly. First, the Medicaid population may be 
quite similar to other groups surveyed about satisfac- 
tion with their current sources of care: people rarely 
admit serious fault with their doctors. ^^ Or Medicaid 
recipients may have low expectations about the kind of 

24. Louis, D. Z. and J. J. McCord, Evaluation of California's 
Prepaid Health Plans (PHPs): Final Report, Contract No. 
HEW-OS-73-194. Santa Barbara, California: General 
Research Corporation, 1974. 

25. Bice, T. W., Enrollment in a Prepaid Group Practice 
(Baltimore: The Johns Hopkins University), 1973. 

26. Lebow, J. L., "Consumer Assessments of Quality of 
Medical Care," Medical Care, 12:4 (April 1974), 328-337. 

23. See Footnote 14. 

medical care available to them." Their satisfaction may 
more appropriately be termed resignation. Bice's 
research at EBMP also suggests that, for Medicaid 
recipients, poor health status is not a good predictor of 
enrollment. At the same time, those who will have less 
out-of-pocket costs by joining an HMO (that is, non- 
Medicaid persons) are more likely to join. It is not un- 
reasonable to assume then, for non-Medicaid persons, 
that some percentage of them will opt for an HMO if 
they anticipate lower health status and consequently 
more health expenditures.^^ While one cannot dismiss 
the potential for an adverse selection of Medicaid 
enrollees, the fact that the HMO cannot usually offer 
reduced out-of-pocket costs to Medicaid enrollees 
should dispel some of this concern. 

One of the least expensive, more effective ways for an 
HMO to attract members is through word-of-mouth. 
Thus, in addition to discovering what advantages of 
HMOs should be emphasized to encourage enrollment, 
policymakers and HMOs might also benefit from identi- 
fying those characteristics of various plans which 
appeal to or alienate their Medicaid members. The ele- 
ments of GHA that enrollees liked the most were con- 
venience, personalized service by nurses, the dental 
program, and quality of care." They reported the most 
dissatisfaction regarding doctors, who were described 
as failing to explain facts and treat them with respect. 

Voluntary disenrollment, of course, is often the expres- 
sion of ultimate dissatisfaction. In the GHA study, only 
2.5 percent voluntarily disenrolled over the 32 months 
of the project. (Eligibility was guaranteed for the entire 
period.) In the EBMP, 25 percent voluntarily disenrolled 
over the 23 month life of the project.^" Disenrollment 
was disproportionately high in the early months of the 
project and the disenrollees had a much higher propor- 
tion of out-of-plan utilization during enrollment. About 
6.5 percent of the low income enrollees in Group 
Health Cooperative under the Seattle Model Cities Pro- 
ject left the plan each year. Overall, a certain per- 
centage of HMO members, usually less than 10 
percent, will exit by choice from the plan each year. 
Medicaid voluntary disenrollment rates are probably 
not different, in general, than the voluntary exit rates of 
employed persons from HMOs. 

27. Owen, E. W., "Proposal to Demonstrate Marketing— Enroll- 
ment and Grievance — Disenrollment Systems," Prepaid 
Health Research, Evaluation and Demonstration Project, 
California Department of Health Services, December 1977. 

28. Bice, T. W., "Risk Vulnerability and Enrollment in a Pre- 
paid Group Practice," Medical Care 13:8 (August 1975) 
pp. 698-703. 

29. See Footnote 13. 

30. Wollstadt, L. J., S. Shapiro, and T. W. Bice, "Disenrollment 
from a Prepaid Group Practice: An Actuarial and Demo- 
graphic Description," Inquiry 15:2 (June 1978), 

pp. 142-150. 


It is impossible to leave the subject of Medicaid and 
HMOs without addressing the California experience. 
What happened in California between 1971 and 1975, 
and the publicity which accompanied it, intimidated 
many other States, Congress, and the Department into 
adopting an extremely cautious approach to the 
development of Medicaid contracts with prepaid health 
plans as well as HMOs. 

In 1971, California initiated major reforms in its Medi- 
cal program, chiefly aimed at controlling runaway 
costs which had risen from $507 million to $1 billion 
between 1967 and 1970. One of the main provisions of 
the Medi-Cal Reform Act of 1971 was the authorization 
of minimal restraints on State contracting with PHPs, 
while placing new restrictions on fee-for-service utiliza- 
tion (that is, prior authorization for use of certain 
services and copayment requirements). PHPs sprung 
up in California overnight, many of them enrolling only 
Medi-Cal recipients. By 1974, 10 percent of the Medi- 
Cal population was receiving services on a prepaid 

Within the first year of the PHP program in California, 
interested and affected constituencies began to com- 
plain about the cost, quality, enrollment practices, and 
corporate accountability of the PHPs which were con- 
tracting with the State. In response, the State enacted 
the Waxman-Duffy Prepaid Health Plan Act of 1973, 
which established some standards of marketing, placed 
a ceiling on the proportion of Medi-Cal enrollees, and 
provided for public hearings on PHP contract renewals. 
Unfortunately, the Waxman-Duffy legislation had little 
impact because the State never effectively enforced the 
law. Scathing articles began to appear in newspapers, 
particularly in Los Angeles. The General Accounting 
Office issued a report which focused on the unusually 
high capitation payments being made to some PHP 
contractors.^' Neighborhood groups, consumer organi- 
zations, welfare workers, public health nurses, legal aid 
societies, local health departments, medical associa- 
tions, and comprehensive health planning councils 
continued to register ever-louder complaints. This 
pressure culminated in hearings on March 13 and 14, 
1975, before Senator Henry Jackson's Subcommittee 
on Investigations of the Committee on Government 
Operations. ^^ During the hearings, sworn testimony 
was heard documenting widespread abuse in the State- 
contracted PHPs: 

31. "Better Controls Needed for Health Maintenance Organi- 
zations Under Medicaid in California," General Accounting 
Office, September 1974. 

32. U.S. Senate, Committee on Government Operations, Per- 
manent Subcommittee on Investigations: Hearings on 
Prepaid Health Plans, March 13-14, 1975, 94th Congress, 
U.S. Government Printing Office, Washington, D.C., 1975. 


(1) quality of care— care unavailable, patients sent to 
county hospitals at expense of county, 
unqualified doctors, refusal to perform necessary 
operations, inferior hospitals providing hospital 
care to PHPs 

(2) fraudulent marketing, enrollment, and disenroll- 
ment practices— deceptive marketing (door-to- 
door enrollers in white coats, telling recipients 
that they would lose their Medi-Cal cards if they 
did not enroll), enrolling derelicts and drug 
addicts merely to get names, forged enrollments, 
involuntary disenrollment when enrollees needed 
expensive care 

(3) complicated networks of for-profit subsidiaries of 
"non-profit" PHPs 

(4) use by PHPs of poor quality, for-profit hospitals — 
The Joint Commission on Accreditation of 
Hospitals came under heavy attack for relying 
solely on physical facilities without the expertise 
to evaluate physician qualifications and quality of 

In addition to documenting the above, the Subcommit- 
tee received sworn affadavits pointing to specific 
examples where the State Department of Health 
Services had quashed investigations, removed overly 
insistent investigators, lost complaint files, failed to 
adequately investigate the backgrounds of applicants 
for contracts, and failed to develop adequate guidelines 
for medical auditors. 

Response at all levels to the California PHP scandal 
has virtually ensured that the situation will not recur. 
The hearings held in March 1975 and again on 
December 14 and 15" resulted in the following 

• The 1976 Amendments to the HMO Act require 
all Medicaid contractors to meet the requirements 
of Title XIII for qualified HMOs, with few 

• DHHS awarded a $5.3 million grant (the Prepaid 
Health Research Evaluation and Demonstration 
Project) to California's Department of Health Ser- 
vices to develop model rate-setting, quality, and 
membership systems for use in California. The 
demonstrations and products of the PHRED pro- 
ject, which are transportable to other States, will 
be discussed in detail at various points throughout 
this report, 

• The Senate enacted new legislation (AB 1963), 
which contains authority to abolish door-to-door 
marketing, mandates actuarial rate-setting, 
improves grievance and disenrollment procedures, 
and mandates accessibility to PHP data. 

33. U.S. Senate, Committee of Government Operations, Per- 
manent Subcommittee on Investigations: Hearings on 
Prepaid Health Plans, December 14 and 15, 1976. 94th 
Congress, U.S. Government Printing Office, Washington, 
D.C., 1977. 


The small number of studies on the Medicare experi- 
ence with HMOs is striking. No study has been com- 
pleted on the utilization of services by Medicare 
beneficiaries enrolled under cost contracts. All studies, 
with the sole exception of the recent Group Health 
Cooperative assessment, are based on pre-1973 
experience. A summary of these studies follows: 

• "Some Aspects of Medicare Experience with 
GPPPs," (Corbin and Krute, 1975): This study 
compares the utilization and reimbursement exper- 
ience of seven GPPPs in New York, Michigan, 
California, and Washington in 1969 and 1970 for 
Medicare enrollees and fee-for-service controls. 
Total reimbursement per beneficiary was lower 
than controls for five of the seven plans. Payments 
for inpatient care were lower than controls in all 
plans, while reimbursement for physicians' services 
was consistently higher. In the two plans where 
lower inpatient costs were offset by higher physi- 
cian costs, the authors suggest this was due 
largely to out-of-plan use. They also suggest that 
the lack of control over hospitals in those plans 
also contributed to their inability to compensate 
for higher physician costs, ^'' 

• "Comparative Costs to the Medicare Program of 
Seven Prepaid Practices and Controls," (Weil, 
1976): The author compares the cost experience of 
Medicare enrollees in seven PHPs in 1969 and 
1970 with similar persons in the fee-for-service 
system, Weil found that physician costs were 
generally higher for the enrollees due to out-of- 
plan use, Costs were also higher for home health 
services, Nevertheless, the net cost effect of enroll- 
ment in 1969 in a PGP was a $69.71 savings per 
enrollee and a $40.20 savings in 1970, The savings 
was achieved primarily through reduced hospitaliza- 
tion, which is shown to be especially effective in 
hospital-based plans. Costs for extended care 
facilities were also less in all PGPs except HIP,^^ 

• "Use of Medicare Benefits Under HIP's 3-Year 
Incentive Reimbursement Experiment," (Densen, 
1978): Between 1970 and 1972, HIP participated in 
an experiment under contract to the Social 
Security Administration (SSA) to demonstrate the 
plan's ability to maintain lower total costs and a 
lesser rate of cost increase for Medicare enrollees 
than the costs incurred by Medicare beneficiaries 
receiving fee-for-service care in the same geo- 
graphic area. The demonstration was evaluated by 
the Harvard Center for Community Health and 
Medical Care, also under SSA contract. The evalu- 
ators found that HIP achieved a lower rate of 

34. Corbin, M. and A. Krute, "Some Aspects of Medicare 
Experience With Group Practice Prepayment Plans," 
Social Security Bulletin. March 1975, pp. 3-11. 

35. Weil, P., "Comparative Costs to the Medicare Program of 
Seven Prepaid Group Practices and Controls," The Mil- 
bank Memorial Fund Quarterly: Health and Society, 
Summer 1976, pp. 339-366. 


increase in total costs between 1969 and 1972 
(27.6 percent for HIP versus 33.1 percent for non- 
HIP) but did not keep total per capita payments 
below those for the non-HIP group. Both HIP and 
non-HIP beneficiaries decreased their use of 
hospital, extended care facilities, and home health 
benefits, while increasing their use of outpatient 
and Medicare Part B services. In fact, the $51 extra 
in physician services used by the HIP group more 
than offset HIP's "savings" under the experiment. 
On average, SSA paid $47 more per capita for 
those enrolled in HIP. Nearly one-third of HIP's 
Medicare enrollees obtained medical services out- 
of-plan in 1972 at a cost of approximately $4.5 mil- 
lion. About 6.8 percent of the Medicare enrollees 
used out-of-plan services exclusively. Only a small 
portion of the total out-of-plan services was not 
covered under HIP's contract. Most of the out-of- 
plan use was linked with hospital use. The HIP 
group also used proportionately more out-of-plan 
specialty physician services than the non-HIP 
population and tended to be more seriously ill than 
those who received all of their services in-plan.^^ 

• "Risk Differential Between Medicare Beneficiaries 
Enrolled and Not Enrolled in an HMO," (Eggers, 
1980): This is a case study which compares the 
utilization and reimbursement experience of Medi- 
care beneficiaries before they joined Group Health 
Cooperative of Puget Sound (between October 
1976 and July 1979) with a group of other Medi- 
care beneficiaries in the Puget Sound area. The 
results suggest that this particular open-enrollment 
population was significantly healthier than the 
comparison group, especially as reflected by their 
pre-enrollment inpatient utilization and reimburse- 
ment. The open-enrollment group had used 52 to 
62 percent fewer days of care per 1,000 than other 
beneficiaries and had 40 to 50 percent less reim- 
bursement. Because of problems with the data, it 
was not possible to calculate utilization and reim- 
bursement amounts for Part B services in a similar 
manner. Therefore Eggers compared the percent- 
age of open enrollees and other Medicare benefici- 
aries who had met the $60 deductible in 1975. 
About 58.6 percent of the comparison group met 
their deductible, while only 52.9 percent of the 
open enrollees had done so. Eggers estimated that 
the open enrollment beneficiaries would have to 
have consumed $440 per user in Part B reimburse- 
ments to offset their lower inpatient reimburse- 

36. Densen, P., E. Jones, I. Altman, and J. Miller, "Use of 
Medicare Benefits Under HIP's S-Vear Incentive Reim- 
bursement Experiment," Final Report, August 1973 (HEW 
Contract No. SSA-70-2905). See also: "Use of Out-of- 
Plan Services by Medicare Members of HIP." Healtti Services 
Research, Fall 1978, p. 243-260. A comprehensive review of 
the HIP results and other Medicare HMO studies is provided 
in "HMOs as an Alternative Mode of Care for the Elderly," 
Karen Lenox, Urban Institute, April, 1978. 

37. Eggers, P., "Risk Differential Between Medicare Benefici- 
aries Enrolled and Not Enrolled in an HMO," Health Care 
Financing Review, Winter 1980, pp. 91-99. 


The methodology and results of this study have been 
featured prominently in HMO reimbursement policy 
discussions since they question a number of premises 
upon which Administration HMO policy is based. 
Among these are: (1) the accuracy of current proce- 
dures for predicting the cost of providing Medicare 
benefits under fee-for-service and for setting Medicare 
HMO reimbursement rates, and (2) the degree to which 
HMOs attract or promote favorable selection. These 
questions and others regarding reimbursement method- 
ology are addressed in detail in other sections of this 


Research on Medicaid eligibles enrolled in HMOs has 
produced no evidence of fundamental differences be- 
tween the Medicaid and non-Medicaid experience. If 
the goal of policymakers is to promote single class 
medicine, this should be encouraging. There remains, 
however, a need for more comparative study of Medi- 
care, Medicaid, and private sector experience in HMOs 
to validate this finding. An interesting place to start 
might be to ascertain whether Medicaid enrollees in 
HMOs like the Harvard Community Health Plan have 
fewer hospital days than other enrollees in this plan, 
and if so, why. Some researchers have argued that 
Medicaid populations have greater need for care than 
the employed.^* 

Available research comparing costs and utilization of 
HMO and non-HMO Medicare and Medicaid benefici- 
aries supports the findings of private sector research 
that HMO members cost less and use fewer hospital 
days. The major unknown factor in this phenomenon is 
the extent to which presumed HMO efficiency is related 
to self-selection of members. In addition, more ques- 
tions need to be answered about out-of-plan utilization. 

Conclusions about the use of ambulatory and preven- 
tive services are highly tentative. Evidence pointing to 
greater or lesser use of ambulatory services by Medi- 
caid recipients can be found. 

Medicare beneficiaries increase their use of ambulatory 
services in general in HMOs. The basis for this finding 
is extremely tentative, however, since only Eggers 
compared the use of services before and after joining 
an HMO. Eggers' study population used less ambula- 
tory care, as well as less inpatient care. 

Accessibility for HMO Medicaid recipients is probably 
improved as far as initial access to a provider, but 
Medicare and Medicaid beneficiaries may face the 
same delays in obtaining appointments as enrolled 
private sector groups. There is some evidence to 
suggest that the process of care for HMO Medicaid 
recipients is better than fee-for-service. However, until 
enrollment incentives other than financial ones are 
understood and acted upon, the likelihood of a 
widespread shift from fee-for-service medicine by 
Federal beneficiaries is highly unlikely. 

38. Aday, L., R. Anderson, and G. Fleming, Health Care in 
the U.S.: Equitable for Whom? (Beverly Hills. California: 
Sage Publications) 1980. 



Current Procedures 

Section 1876 of the Social Security Act authorizes two 
Medicare reimbursement options: cost and risk. Cost 
reimbursement is essentially an application of the 
Medicare cost principles of reimbursement, as specified 
in the Provider Reimbursement Manual (although cer- 
tain exceptions are permitted). The HMO is reimbursed 
on an interim per capita basis, which is established 
from an annual operation budget and enrollment pro- 
jections. Interim estimated cost reports and enrollment 
projections are submitted on a quarterly basis and are 
used to adjust the interim per capita rate. A final 
audited cost report is subsequently submitted by the 
HMO. There is no requirement under cost reimburse- 
ment that the HMO's per capita costs should be less 
than the equivalent fee for service costs. 

Risk reimbursement introduces the concept of the 
adjusted average per capita cost (AAPCC), which is 
defined in Section 1876: 

"The term "adjusted average per capita cost" means 
the average per capita amount that the Secretary 
determines (on the basis of actual experience, or 
retrospective actuarial equivalent based on an ade- 
quate sample and other information and data, in the 
geographic area served by a health maintenance 
organization or in a similar area served by a health 
maintenance organization with appropriate adjust- 
ment to assure actuarial equivalence, including 
adjustments relating to age distribution, sex, race, 
institutional status, disability status, and any other 
relevant factors) would be payable in any contract 
year for services covered under this title and types 
of expenses otherwise reimbursable under this title 
(including administrative costs incurred by organiza- 
tions described in sections 1816 and 1842) if such 
services were to be furnished by other than such 
health maintenance organization." 

Under risk reimbursement, the HMO is reimbursed its 
cost, and if the costs are under the AAPCC, the HMO 
receives 50 percent of the savings (defined as the dif- 
ference between the AAPCC and cost), up to 10 percent 
of the AAPCC. If the costs are higher than the AAPCC, 
the HMO must absorb the losses, but these can be 
carried forward into subsequent years and can offset 
future savings. The current risk reimbursement is also 
linked to the determination of costs following Medicare 
principles of cost determination. The HMO is therefore 
unable to know, at the beginning of the service year, 
the expected revenue from its Medicare enrollees. This 
is contrary to the procedures followed by HMOs in the 
private sector, where they are on full risk, receive a 
fixed premium, and can appropriately plan and budget. 
Currently only one HMO, the Group Health Coopera- 
tive of Puget Sound, is reimbursed on a risk basis. 

The key concept under the current risk reimbursement 
and proposed alternative full risk procedures, is the 
AAPCC. The AAPCC measures what it would have cost 
the Medicare program if the individuals enrolled in the 
HMO had received benefits in the fee-for-service sec- 
tor. To calculate the AAPCC, HCFA's Office of the 
Actuary (OACT) does the following: 

(1) determines the U.S. per capita cost 

(2) determines the service area per capita cost by 
multiplying by the appropriate service area geo- 
graphic and age-sex index (The service area 
represents the counties which the HMO serves.) 

(3) adjusts the service area per capita costs by the 
appropriate age-sex underwriting indices to 
obtain HMO specific cost. (These indices adjust 
for differences between the service area and 
HMO age, sex, institutional status and welfare 

The AAPCC thus starts with national costs, then 
applies appropriate indices to obtain county-specific 
data, and finally applies the age-sex underwriting 
indices to obtain HMO specific cost. 

This procedure for calculating the AAPCC enables the 
HCFA actuaries to calculate the AAPCC for any county 
in the U.S. The methodology used to calculate the 
AAPCC was reviewed by a special panel of actuaries 
convened by HCFA in January 1979. The report pre- 
pared by the panel stated:^' 

"We believe the AAPCC computation responds well 
to the legislative requirements in Section 1876 and 
that the AAPCC formula contains a sufficient level of 
accuracy. However, there was expressed concern 
about the statistical validity of the geographic factors 
and a suggestion that the Underwriting Index might 
be further validated." 

There are some limitations which must be recognized if 
the AAPCC is adopted as the basis for new risk reim- 
bursement procedures: 

(1) The age-sex underwriting indices are based on 
data collected from three years of the Current 
Medicare Survey (CMS). CMS data were self- 
reported, with all the limitations this imposes, and 
are no longer collected. 

(2) The finest level at which the AAPCC can be calcu- 
lated is the county, since the age-sex and geo- 
graphic indices are determined at this level. 
Although the capability exists theoretically to 
calculate these indices on a zip code level, this 
would require extensive computer runs. Zip code- 
based indices would undoubtedly be more 
unstable than county-based indices. On the other 
hand, if an HMO's service area is not representive 
of the county in which it is located (a problem 
likely to occur in counties with large cities or dis- 
tinct urban and rural sectors), serious errors in the 
AAPCC calculation mav occur. 

39. Report of the Medicare HMO Actuarial Review Panel for 
the Office of Reimbursement Practices. HCFA. January 


(3) The calculation of the AAPCC requires placing all 
Medicare individuals in the service area or the 
HMO, by age and sex, in the follov\/ing separate 
cells: (1) institutionalized, (2) welfare and not 
institutionalized, and (3) all others. The counts for 
these cells are established from different data 
sources which identify the number of individuals 
who are institutionalized, the number who are on 
welfare, and all others. The calculation requires 
the assumption that all institutionalized individuals 
are on welfare. The validity of this assumption 
needs to be verified. 

(4) The welfare adjustment indices are based on 
national CMS data. The wide variation in Medicaid 
coverage across States may introduce an 
unknown bias. 

(5) The age-sex underwriting indices have not been 
redetermined since they were calculated in 1977. 
There has been no assessment of their stability 
over time. 

(6) If an HMO enrolls a significantly large proportion 
of its service area, it may not be appropriate to 
use that county to calculate an AAPCC. A neigh- 
boring county may have to be used, but this raises 
the issue of how to establish that the characteris- 
tics of the counties are reasonably matched. 

(7) The calculation of an AAPCC currently requires 
approximately five weeks— two for the demo- 
graphic data to be obtained from the computer 
files and three for OACT to perform the calcula- 
tions. If a large number of HMOs enroll Medicare 
beneficiaries under a risk arrangement, this pro- 
cess must be expedited. Since most HMOs will 
operate on a calendar year basis, most AAPCC 
calculations are required by October (three 
months prior to the new service year, permitting 
the HMO to calculate and communicate new 
premiums). If institutional data must be obtained, 
they may have to be collected as early as March 
or April. 

In addition to these limitations of the AAPCC, there are 
some administrative problems which arise in collecting 
the data necessary to determine the AAPCC. 

(1) Welfare status is determined by an indicator in the 
Health Insurance Master Beneficiary File that 
premiums are paid by Medicaid. Although most 
States have buy-in agreements, four did not, as of 
March 1980: Louisiana, Alaska, Oregon, and 
Wyoming. Data then have to be obtained directly 
from the State, which may be difficult. 

(2) The major administrative problem is collecting 
accurate data on institutionalization. The defini- 
tion of institutionalized individuals used by the 
HCFA actuaries is: "Medicare beneficiaries residing 
for over 30 days in a nursing home, sanitarium, 
rest home, convalescent home or long-term care 
hospital." Since the range of institutions included 
in this definition is considerably broader than the 
skilled nursing facility (SNF) care covered by 
Medicare, the number of institutionalized Medicare 
beneficiaries cannot be obtained from HCFA files. 
Independent data sources, such as State Medicaid 
programs or independent surveys, must be used 

to obtain this data. A discussion on how institu- 
tional data were obtained for IKe demonstration 


sites and alternative strategies which could be 
used in the future can be found in Appendix B. 

Determination of Prospective AAPCC 

Establishing a procedure for calculating a prospective 
AAPCC is relatively straightforward and follows the 
procedure used to calculate the retrospective AAPCC. 
In the simplest approach, the value of all parameters 
used to calculate the retrospective AAPCC must be 
projected forward to estimate their values during the 
prospective service year. The HCFA Office of the 
Actuary has established the following procedures for 

(1) Calculate the U.S. per capita cost with the most 
current projection used by OACT in preparing 
the annual reports to Congress on the Medicare 
Trust Funds. 

(2) Use a five year unweighted average for the geo- 
graphic index. 

(3) Use a five year unweighted average of the age- 
sex index. 

(4) Assume that the age-sex underwriting indices 
remain constant over time. 

(5) Obtain the most current data on service area 
Medicare enrollment (that is, individuals on wel- 
fare, institutionalized, or other) prior to calcu- 
lating the prospective AAPCC, and assume that 
the data remain constant during the service year. 

Using the above procedures, a prospective AAPCC can 
be calculated. However, because of the sensitivity of 
the issue of favorable or unfavorable selection (which 
will be discussed in greater detail later), OACT believes 
it is not appropriate to project enrollment statistics, 
unless no other alternative can be found. This has 
resulted in the development of the rate book approach, 
in which OACT calculates the appropriate reimburse- 
ment for each cell for which an age-sex underwriting 
index has been established. The HMO is thus reim- 
bursed a different rate for enrolling a 65 year old male 
than for a 75 year old male or a female. Rates for 15 
rate cells for each county in the HMO's service area 
must be calculated for both Part A and Part B. While 
this procedure clearly eliminates the selectivity issue 
from both the HMO's and the government's viewpoint, 
it introduces some other problems: 

(1) The HMO is not able to accurately project its 
revenue from enrolling a fixed number of enrollees, 
since a different rate is established depending on 
the enrollees' characteristics. (However with pay- 
ment based on these characteristics, the HMO 
assumes less risk.) 

(2) The HMO must establish procedures to identify 
when a beneficiary becomes institutionalized, 
and report this as part of its accretion report to 
HCFA. Age, welfare status, and county of resi- 
dence are determined from the Health Insurance 
Master Beneficiary Files. For those States with- 
out buy-in agreements, the HMO would have to 
be able to determine the welfare status of its 
enrollees. This may require HMOs to make sys- 
tems modifications to accommodate Medicare 

The rate book approach is only being implemented in 
the InterStudy demonstration. (See Appendix A for a 
brief description of the Medicare capitation demonstra- 
tions.) Kaiser, Portland negotiated a hybrid approach 
during the developmental phase of its demonstration. 
Kaiser would not accept the rate book approach 
because it would have required changes in their data 
processing procedure, and because they were unable 
to report the welfare and institutional data on a 
monthly basis. (Oregon is one of four States with no 
buy-in.) Since Kaiser did have age-sex distribution data 
on their current group prepaid practice enrollees, it 
was agreed, prior to the initiation of the service 
contract, to assume an enrollment distribution to 
calculate one interim AAPCC, which would be used to 
reimburse Kaiser for each enrollee. During the year, 
Kaiser would obtain actual counts of the age and sex 
mix of the enrollees, and HCFA would apply a one 
month count of welfare and institutionalized individuals 
for the full calendar year. These data are used to make 
a retroactive adjustment to the payment made based 
on the interim AAPCC. 

Once an HMO has a stable enrollment, the differences 
between using one projected AAPCC, the hybrid 
approach, or the rate book will be minimal. During the 
initial enrollment build-up, errors incorrectly projecting 
an HMO enrollment mix can have a significant impact 
on the AAPCC, since the age-sex underwriting factors 
vary so widely. 

Any problems in accuracy or bias in calculating the 
AAPCC become more sensitive when calculating a 
prospective AAPCC. One final example will show how 
difficult it is to establish a procedure that will be 
accurate under ail possible circumstances. Consider 
the situation where an HMO is established in a medi- 
cally underserved area. Because of access problems. 
Medicare beneficiaries may not be using the full range 
of services to which they are entitled. This would be 
reflected in a low AAPCC. For those beneficiaries en- 
rolling in the HMO, these access barriers are removed, 
but the HMO's reimbursement is still limited by the 

Experimental Reimbursement Procedures 

The demonstration projects will test three alternative 
risk reimbursements: (1) percentage of the AAPCC, (2) 
adjusted community rate, and (3) actuarial. None of 
these approaches is linked to cost reimbursement. The 
AAPCC is used to limit the reimbursement rate of the 
latter two methodologies. 

None of the demonstration projects exactly duplicates 
the 1980 Administration proposal (not adopted by the 
96th Congress) which operated as follows: 

(1) The HMO is reimbursed 95 percent of the 
AAPCC. The difference between this figure and 
the adjusted community rate (ACR) is defined as 

(2) These savings are used by the HMO to provide 
additional services or reduced premiums. Previ- 
ous legislative proposals have listed a specific 
order in which savings must be applied (preven- 
tive services, reduced premiums, and additional 

services); the 1979 House bill eliminated the 
order and left the decision to the HMO. 

The ACR calculation is based on adjusting the HMO's 
utilization and cost statistics in the private sector for 
the characteristics of the Medicare population. For an 
HMO with an enrolled Medicare population (through a 
cost contract or as a GPPP), the determination of the 
ACR is essentially straightforward. Adjustment factors 
are established by calculating the ratio of Medicare 
utilization within the HMO to private sector utilization 
for a specific service. (This is known as a volume 
factor.) For physician services, current HMO regula- 
tions permit the application of an additional 20 percent 
time and complexity factor. If these data are not avail- 
able, the establishment of adjustment factors becomes 
more controversial and their accuracy problematic. The 
key factors that must be established are for hospital 
and physician utilization. The following procedures 
have been used: 

(1) Assume that factors available from other HMOs 
are applicable. Kaiser, Portland has extensively 
published utilization data for their populations. A 
formula that can be applied is: 

HMO Medicare Rate = HMO Private Rate x Kaiser Medicare Rate 

Kaiser Private Rate 

Many HMOs are reluctant, however, to assume 
that they will initially be as efficient with a new 
Medicare population as well-established HMOs. 

(2) Obtain data on the fee-for-service utilization 
experience in the service area and make appro- 
priate adjustments. It is usually assumed that 
hospital days will be reduced by 5 to 25 percent, 
while physician services will be marginally 
increased. For the demonstration sites, HCFA 
provided Medicare utilization data by age and 
sex for each county in the HMO's service area. 

(3) An actuarial firm will provide the HMO with utili- 
zation data based on "standard file formula" from 
the actuary. 

The actuarial model is closely related to the ACR pro- 
cedures, and sometimes the distinction is not clear in 
practice. Theoretically, the actuarial model builds up 
the rate from scratch, without linking it to the HMO's 
experience with the private sector. The problems of 
establishing HMO Medicare utilization statistics are 
essentially the same as those described above. 

A controversial issue in setting Medicare HMO rates is 
how administrative costs are distributed among the 
Medicare program and the private sector. For example, 
under HMO cost reimbursement, marketing costs are 
allocated by member-months, although it is generally 
recognized that the Medicare (and Medicaid) popula- 
tion is more difficult and expensive to market. Large 
concentrations of Medicare beneficiaries, as found in 
private business, are not available for marketing. The 
ACR determination provides flexibility in adjusting for 
marketing costs. Use of the adjusted community rate ' 
can also reduce controversy concerning what are 
appropriate allowances for a retention factor for 
earnings and reserves. The adjusted community rate 
concept permits the application of the identical 
retention factor for the Medicare population. 


A final issue which may arise in calculating the 
adjusted community rate is: Which community rate 
should be used? Large HMOs may offer many different 
rates which could be considered the community rate, 
and it is not always obvious which is the appropriate 
rate to adjust Detailed, supported justification for com- 
ponents of the community rate may not be available. If 
available, the rate and supporting documentation used 
by the HMO to establish its rate for the Federal 
Employees Health Benefit Program provides an inde- 
pendent benchmark which can be used as the 
community rate. 

The rate book approach for the AAPCC was estab- 
lished to avoid uncertainty in the HMO's actual enroll- 
ment. This uncertainty cannot be avoided, however, in 
establishing the ACR. To prospectively establish the 
ACR and estimated savings, the enrollment distribution 
must be assumed initially. Fortunately, serious errors 
should occur only for an HMO with no prior Medicare 
enrollment, and then only for the initial years. Further- 
more, the error should only affect the determination of 
additional benefits that the Medicare beneficiary will 
receive, since the rate book approach will assume 
proper payment to the HMO. 

Limited experience with the demonstration projects indi- 
cates that the proposal to set rates at 95 percent of the 
AAPCC has the strong advantage of being relatively 
straightforward and minimizing HCFA's review process. 
The actuarial approach requires extensive review and 
justification and becomes similar to a budget review 
process. The disadvantage lies in the apparent arbitrari- 
ness of the percentage actually selected. Why not 90 per- 
cent of the AAPCC, or even 80 percent, rather than 95 

During the developmental phases of the demonstration 
contracts, great variability was displayed in the ability 
of the contractors to come under the 95 percent cri- 
terion. Two of the plans were able to provide significant 
additional benefits while being reimbursed at 90 to 95 
percent of the AAPCC. One plan was reimbursed at 98 
to 99 percent of the AAPCC, with limited additional 
benefits. The other participating projects will be reim- 
bursed at 95 percent of the AAPCC and will provide a 
limited package of additional benefits. One benefit 
which all plans provide, but which is not always recog- 
nized as an additional benefit, is that all Part B physi- 
cians services are provided as if on assignment. There 
are no additional, out-of-pocket costs for these 
services. This can represent a significant savings to the 
Medicare beneficiary, since approximately 50 percent 
of claims are currently not accepted on assignment. 
Two projects were not able to establish rates below the 
AAPCC and were therefore not permitted to continue 
into the implementation phase. 

Selection Issue 

The most controversial issue in evaluating the ade- 
quacy of reimbursement rates for HMOs is favorable/ 
adverse selection. There has been much debate on this 
matter, with the available studies providing ambiguous 
results. On a theoretical basis, convincing arguments 
can be made for both possibilities. Favorable selection 
should occur because individuals who already have 
ties with physicians (and presumably have poorer 
health status than the average) will not join HMOs 
unless they are dissatisfied with their current practi- 
tioners. On the other hand, one can hypothesize that 
individuals who anticipate high medical bills are most 
likely to change from the fee-for-service sector (with 
potential high out-of-pocket costs) to an HMO setting, 
with the knowledge that for a fixed premium, all their 
health costs will be covered. 

In the private sector, the debate concerning favorable/ 
adverse selection is primarily of academic or policy 
interest, since the HMO is permitted to set its own 
community rate, subject only to the competitive forces 
working in the marketplace. Except for specific limited 
conditions, a Federally-qualified HMO is not permitted 
to make any demographic adjustments to its commun- 
ity rate because of the special characteristics of its 
enrolled population. The HMO may reduce some possi- 
bility of adverse selection by health screening and by 
assuming responsibility for an individual only after dis- 
charge from an institution. 

When rates are established by a government agency 
rather than by the HMO, the selection issue intensifies. 
The AAPCC should reduce some of the controversy in 
that it attempts to adjust for the specific demographic 
characteristics of HMO enrollees by controlling for 
their age, sex, welfare, and institutional status. Thus, 
even if the HMO enrolled a significantly younger (or 
older) population than is present in the fee-for-service 
sector, the AAPCC should make the proper 

The AAPCC would not, however, adjust for differences 
in health status within a specific demographic group. 
For example, in the study, "Risk Differential Between 
Medicare Beneficiaries Enrolled and Not Enrolled in an 
HMO," described above, Eggers examined the selection 
issue for Medicare open enrollees in the Group Health 
Cooperative of Puget Sound. The study compared the 
inpatient hospital use and associated costs by Medi- 
care beneficiaries for the two years prior to joining 
GHC to the inpatient utilization and costs of Medicare 
beneficiaries who resided in the same geographic area 
for the same time period but who did not enroll in the 
HMO. Specifically, the study compared the Medicare 
claims of 887 beneficiaries who later joined GHC 
during one of its open enrollment periods to claims 
generated by approximately 200,000 beneficiaries who 
did not join GHC. Eggers concluded that during the 


two years prior to their enrollment in GHC, the GHC 
enrollees' hospital inpatient utilization was 52 to 62 
percent lower than that of the connparison group, after 
adjustments had been made for age, sex, welfare, and 
institutional status. The per capita hospital costs for 
those same inpatient services were 40 to 53 percent 
below the comparable population. Also, the open 
enrollees' use of Part B services appeared to be lower, 
as measured by the percent of people fulfilling the Part 
B deductible. 

The Eggers study represents a methodogical approach 
to the selection issue which can be replicated in future 
studies. Comparing prior utilization of HMO enrollees 
to a fee-for-service control population is a much 
sounder approach than simply looking at age-sex dis- 
tribution or utilization after enrollment, as was done in 
most previous studies. The results indicate that further 
analysis of the AAPCC methodology may be warranted. 
It is premature to generalize from the results of only 
one study. HCFA's Office of Research does plan to 
refine the methodology and to undertake a similar 
study for the current capitation demonstrations, which 
will offer a larger population base. It may also be use- 
ful to conduct a similar study with Medicare cost con- 
tracts which have open enrollment periods. 

Should the Eggers results be verified, adjustment for 
health status in the reimbursement methodology or 
modification of enrollment procedures would be neces- 
sary. Some have argued that HCFA has not properly 
implemented the Section 1876 statutory requirement to 
consider disability status as a factor in determining the 
AAPCC. HCFA has interpreted this requirement as 
requiring the calculation of a separate AAPCC for the 
disabled and aged populations. Opponents of this view 
believe that the legislative intent was to consider health 
status as one of the actuarial adjustments. 

It would be extremely difficult to incorporate health 
status in the AAPCC since there is no generally accept- 
able measure of health status. Even if such a measure 
existed, it could not simply be incorporated in the 
AAPCC because there are no currently available or 
planned Medicare national data bases which include an 
acceptable measure of health status and the currently 
used factors. It would take at least three to five years 
of developmental work to incorporate a health status 
measure in the AAPCC. 

Another alternative is to address the issue of selection 
administratively. If one cannot incorporate health 
status in the AAPCC, one could closely monitor the 
open enrollment procedure to reduce the possibility of 
selection. The Eggers methodology could be used to 
compare prior hospital utilization of the HMO enrollees 
and a comparison group to monitor the open enroll- 
ment procedure. While this could be accomplished 
rather easily, it does not address the administrative 
action to be taken if either markedly favorable or 
unfavorable selection occurred. 

A third approach might be to use a community rate, as 
used in the private sector. The community rate does 
not permit any adjustments for demographic character- 
istics, yet HMOs are generally flourishing. (The cynic 
would argue that this indicates that HMOs are experi- 
encing favorable selection.) If a number of HMOs can 
demonstrate financial stability by enrolling the Medi- 
care population on a risk basis using an AAPCC or 
community rating methodology, other HMOs will 
follow. (Of course, the cynic will then argue that this 
only proves that the AAPCC is biased toward the 
HMO.) The basic issue boils down to how complicated 
a risk reimbursement methodology one wishes to 
develop to minimize unknown factors. The only level at 
which one can eliminate all risk of overpayment is cost 
reimbursement, with all its attendant inflationary im- 
pulses. However, if the government wishes to 
encourage HMO contracts under Medicare, it may have 
to adopt a straightforward procedure for establishing 
the AAPCC."" 


Current Procedures 

On May 9, 1975, the Department published HMO regu- 
lations which included requirements for establishing 
Medicaid capitation rates. The regulations require: 

• that the State agency specify the actuarial basis 
for computation of the premium rate or subscrip- 
tion charge specified in the contract; 

• that these rates or charges be reasonable; and 

• that the State pay no more than the equivalent per 
capita cost that would have been paid on a fee-for- 
service basis. 

Two general approaches are being used to meet these 
broad requirements: fixed percentage of the equivalent 
fee-for-service costs and the actuarial approach. The 
fee-for-service percentage is attractive to States 
because it is administratively straightforward and 
assures a savings over expenditures which theoretically 
would have been incurred had the Medicaid recipient 
not been enrolled in the HMO. The fee-for-service per- 
centage can also result in savings for HMOs, particu- 
larly those which achieve the reductions in inpatient 
utilization typically realized by efficient plans. By the 
same token, some view such potential for savings as 
too susceptible to under- or overpayments to the 

40. The opposite viewpoint is presented in a recent Urban 
Institute publication, National i-iealtti Insurance: Conflict- 
ing Goals and Policy Choices, edited by Judith Fader, 
John Holahan, and Theodore Marmor. The authors argue 
that "present methods of calculations of the adjusted 
average per capita cost are extremely crude. Other vari- 
ables [besides those currently used] likely to be highly 
correlated with the elderly's use of health services would 
include the presence and kind of chronic illness, Medi- 
caid eligibility, availability of private insurance, conversion 
and open enrollment status, income, education, marital 
status, and living arrangement." 


The chief advantage of actuarial rate-setting is that the 
results should mitigate the potential for adverse selec- 
tion risk to the HMOs and favorable selection risk to 
the State. Hov^ever, actuarial rate-setting requires 
front-end development of a reliable HMO information 
system which will not, by its own administrative 
burden, discourage plans from contracting. Moreover, 
the State must have the technical expertise to apply 
actuarial rate-setting methods. 

PHRED and the Evolution of Rate-setting in California 

In the early years of the PHP program in California, the 
State negotiated rates independently with each plan. 
Negotiators used 90 percent of the fee-for-service per 
capita equivalent costs as the overall target but per- 
mitted considerable variation among the plans. Some 
of the rates even exceeded the fee-for-service equiva- 
lent costs. In the aggregate, the plans' rates were in the 
range of 83 to 85 percent of fee-for-service rates. Cali- 
fornia responded to overall dissatisfaction with this 
rate-setting approach in 1974 by developing standard 
rates for the four aid categories for each county. The 
method essentially was to apply the same percentage 
increase to PHP rates as was projected for the fee-for- 
service budget. But this did not mollify the critics. The 
PHPs, State and Federal auditors, and Congress called 
in unison for the State to adopt a rate-setting method 
which incorporated contemporary standards of 
actuarial analysis. 

Recognizing the difficulties in this, HCFA supported 
the Prepaid Health Research, Evaluation, and Demon- 
stration (PHRED) project in California, which had as a 
major goal the development of an actuarially sound 
rate-setting methodology to be implemented by the 
State. PHRED staff were also asked to analyze the ad- 
verse selection issue which had arisen in connection 
with the capitation payment to the Foundation Commun- 
ity Health Plan (FCHP). 

Because of delays in implementing the project, it was 
impossible for PHRED to undertake a small scale 
demonstration and have it used by the State in time for 
the fiscal year 1977-78 rates. Instead, PHRED 
attempted to apply actuarial methods to set rates for all 
PHPs, while at the same time, the State analysts repli- 
cated their traditional method of projecting increases. 
The statistical reports being submitted at that time by 
the plans aggregated all four aid categories. Because 
these reports formed the major source of data for rate- 
setting, and because AFDC represented 87 percent of 
PHP enrollees and 83 percent of the capitation pay- 
ments, PHRED was able to calculate fiscal year 1977- 
78 actuarial rates for AFDC only. 

PHRED's actuarial methods built upon the recommen- 
dations of an earlier series of reports prepared by the 
Martin Segal Company for the State. The model, 
developed from these recommendations, considers 
three principal factors: demographics, utilization, and 
costs. The ironic outcome of this rate-setting demon- 
stration was that the aggregate actuarial AFDC rate 
was $34.84, and the traditionally set rate v/as $34.89— 
both about 83 percent of fee-for-service. PHRED offers 


two explanations for this parity: (1) the possibility that 
global assumptions suffice in actuarial work related to 
health care costs, and (2) that the 83 percent of the 
fee-for-service rate set arbitrarily in earlier years turned 
into a self-fulfilling prophecy for future costs and 

A major product of PHRED under this component of 
the project was their "Ratesetting Guide for Prepaid 
Medicaid Contracts." This guide is designed to assist 
Medicaid agencies with the policy and technical issues 
of actuarial rate-setting for prepaid Medicaid contracts. 
The guide includes chapters on: 

• the conceptual model 

• determining the fee-for-service maximum 

• determining the community rate maximum 

• other actuarial rate-setting methods (that is, cost 
modeling, budget modeling, and community rate 

• adjustment factors 

• risk sharing 

• administration 

For the fiscal year 1978-79 rates, the California Depart- 
ment of Health Services used the PHRED model for all 
four aid categories. This was largely facilitated by the 
passage in September 1977 of a State law (AB 1693), 
mandating accessibility to PHP data and empowering 
the State to specify what data it would collect. The law 
also requires the California Department of Health 
Services to use actuarial rate-setting methods and 
specifies age, sex, and aid category as minimum 
demographic characteristics that must be included in 
such methods. That year's rates set by the Department 
were developed in the following steps: 

(1) The State used experience data reported by PHPs 
to develop utilization and cost assumptions. 

(2) Cost assumptions were projected in accordance 
with current Consumer Price Index trends and 
Medi-Cal fee-for-service trends. 

(3) Increased PHP responsibility for mental health 
services and child screening increased the rates; 
they were reduced for "other insurance" recovery. 

(4) For the adult aid categories (ATD, OAA, AB), 
adjustments were made for Medicare coverage. 

(5) For AFDC, the standard rates were adjusted to the 
utilization experience of each PHP (that is, an 
age-sex adjustment was applied). 

(6) For ATD and OAA, the standard rates were ad- 
justed to the utilization experience of each PHP. 

(7) For ATD and OAA, a risk-sharing rule was 

(8) To reflect the impact of reduced abortion 
coverage, adjustments were made for risk-sharing 
related to possible delivery rate increases. 

(9) Resulting rates were compared with fee-for- 
service per capita costs and with PHPs' commer- 
cial rates. 

41. For a detailed discussion about risk-sharing, see the 
article written by the former PHRED project director, 
Rigby Leighton: "Risk Sharing: A New Reimbursement 
Alternative" (in Perspectives on Medicaid and Medicare 
Management, September 1979, HCFA-79-20021). In addi- 
tion to outlining general principals of risk-sharing, the 
paper describes a form of selective risk-sharing being 
used by California. 

As a result of that year's rate analyses, the State dis- 
covered that the rates previously paid for the ATD. 
OAA, and AB aid categories were significantly higher 
than could be justified on the basis of actual PHP 
utilization in these categories. Since this finding cast 
doubt upon the validity of the actuarial assumptions, a 
risk-sharing formula for these categories was 
developed for the 1978-79 rates: The plan and the State 
shared risk for cost effects of increases in utilization 
between 100 percent and 150 percent of the plan's 
experience-based utilization assumptions. Alt costs in 
excess of 150 percent v/ere borne by the State to a 
maximum of the fiscal year 1977-78 aggregate rates. In 
effect, then, the plan is held accountable only for those 
cost effects over which ii has direct control. 

The Department of Health Services ran into opposition 
from the State Department of Finance in gaining 
approval of the rates developed as described above. 
The Department of Finance opposed actuarial rate- 
setting primarily on the grounds that tying rates to cost 
and utilization experience put the State in the same 
position with HMOs as it was in with other providers 
receiving cost reimbursement. The Departments ulti- 
mately reached a compromise in which rates were 
derived by the traditional projection method and 
adjusted for AFDC by PHRED's age sex factors. 

The PHRED project also examined the adverse selec- 
tion issue in an IPA setting. During the 1972-76 period. 
the Foundation Community Health Plan (FCHP) was an 
IPA model PHP contracting with the State of California. 
The capitation rates paid by the State to FCHP were 
unusually high. In 1974, the FCHP capitation rates 
were approximately 16 percent higher than the esti- 
mated per capita costs for the non-enrolled Medi-Cal 
beneficiaries in the five county area served. In contrast, 
the capitation rates for the other PHPs averaged about 
15 percent less than fee-for-service per capita costs. 

The State and FCHP originally justified the higher rates 
by arguing that FCHP experienced adverse selection. 
California responded as follows: 

■'The Department of Health allowed higher rates 
because the FCHP enrollees include a disproportion- 
ate number of sick people. The Department con- 
cluded that this was the case because the FCHP 
made enrollments through physician offices and 
those enrolled were, therefore, in more need of 
medical care."-' 

This is the crux of the argument often made by IPAs of 
why they experience adverse selection. 

When HEW declared an audit exception because of the 
alleged overpayment, the PHRED project analyzed the 
utilization experience of FCHP enrollees and fee-for- 
service beneficiaries in the five county Sacramento 
area. (Regulations do not permit capitation payments 
to HMOs higher than the fee-for-service maximum.) 

The PHRED report first examined the claim that enroll- 
ment through physician offices had resulted in adverse 
selection and discovered the claim could not be 
supported. The great majority of individuals had 
enrolled in FCHP prior to the effort to have member 
physicians promote enrollment. The tenuousness of 
this adverse selection argument is further supported by 
PHRED's reasoning as to why adverse selection may 
have occurred: 

'Our conclusion is that the "enrolling through the 
doctor's office" image is specious, a convenient myth 
that became increasingly credible through repetition. 
The irony of the situation is that FCHP may have ex- 
perienced adverse selection for exactly the opposite 
reason — the organization undoubtedly assigned a 
primary care physician to a significant number of 
individuals who previously did not have a personal 
physician. (California requires that each PHP enrollee 
be assigned to a primary care physician. i Each sucn 
individual thereby had a higher degree of access not 
only to physician services but to all those health care 
services for which a physician is the 'gate keeper." " 

In testing this hypothesis, the PHRED project did not 
have access to utilization experience of FCHP 
enrollees and a control group prior to enrollment las 
was used in the Eggers study) and had to rely on 
actual utilization subsequent to enrollment. A simple 
comparison of age and sex indicated no difference 
betyt-een the two groups. lAge-sex adjustments consti- 
tute the only actuarial adjustment used by the State in 
establishing HMO rates.) Since differences in utilization 
could have been caused by either favorable unfavor- 
able selection or variations in utilization control mech- 
anisms, the PHRED analysis attempted to control for 
the latter variable and hence assign any utilization dif- 
ferences to the selection factor. 

This was accomplished through two key assumptions: 

(1) Acute hospitalization— the most expensive single 
category of service— was subject to the same 
utilization control, the Certified Hospital Admis- 
sion Program. 

(2) FCHP controlled utilization for other services 
similar to those used by the State fiscal 

The analysis of utilization patterns indicated that, for 
the AFDC population, adverse selection occurred in 
terms of ambulatory services, but favorable selection 
occurred in terms of acute hospital care. For ATD. the 
utilization pattern showed substantial adverse selection 
for both ambulatory and acute hospital care. Adverse 
selection was defined as increased utilization of services. 

The PHRED analysis presents a careful methodology 
for analysis of the selection issue based on actual utili- 
zation in the HMO. but the validity of the results can 
only be accepted if the key assumptions are confirmed. 
The Eggers methodology eliminates the necessity of 
making these assumptions. 

42. 'The FCHP Audit Exception: Evaluation of the Adverse 
Selection Hypothesis." PHRED Report, July 1977. 




There are two primary problems in marketing HMOs to 
Medicare beneficiaries. Identifying the Medicare popu- 
lation presents the first obstacle. Many HMOs believe 
that direct mailing of literature to Medicare benefici- 
aries would be effective, but HCFA is precluded by the 
Privacy Act from furnishing lists to HMOs of Medicare 
beneficiaries' names and addresses. To partially over- 
come this problem, HCFA's Group Health Plan 
Operations Staff makes direct mailings to beneficiaries 
in HMO service areas, notifying them in standard lan- 
guage that there are one or more HMOs in the area 
that serve Medicare beneficiaries. The message does 
not explicitly advocate HMO membership, and it does 
not give details on the specific HMO's characteristics. 

The second problem in marketing to Medicare benefi- 
ciaries is cost. Even after the population has been 
identified, perhaps at considerable cost, individuals 
must be reached. Unlike the private sector, there are 
generally no employer groups or unions through which 
information can be conveniently disseminated. Often, 
the HMO must resort to one-on-one presentations. 
Current Medicare cost-based reimbursement mech- 
anisms provide for some reimbursement of marketing 
costs, but the amount is inadequate in many cases. 
Under Medicare reimbursement principles, HMOs are 
allowed to allocate marketing costs among programs 
on an enrollee basis, but this does not reflect the 
relatively greater cost of marketing to the Medicare 
population. HMOs thus have little incentive to market 

Demonstration Marlteting Procedures and Resuits 

Community Health Plan obtained a mailing list from 
Blue Cross of Massachusetts of all Medicare supple- 
mental policyholders, which they used to disseminate 
literature in Worcester County, Massachusetts. How- 
ever, Fallon's main effort focused on its "health fair" at 
the Fallon Clinic held prior to its enrollment campaign. 
The health fair offered descriptions of the Senior Plan 
to potential enrollees, tours of the Fallon facilities, and 
several booths and exhibits dealing with a variety of 
topics, including safety, retirement, and chronic ill- 
nesses (for example, diabetes). 

Approximately 4,000 people attended the health fair. In 
addition, Fallon held five open houses, attended by a 
total of 400 to 500 people, in which talks on the Senior 
Plan were given and tours of the Fallon facilities were 
arranged. Fallon also placed newspaper ads in two 
major newspapers and three small local ones. The ads 
contained a card that interested readers could send in 
to receive a brochure and application. Fallon received 
over 3,000 returned cards. 

Fallon began its marketing efforts on February 7, 1980 
and within a month had met its first year projection of 
3,600 enrollees. (its open enrollment period was ter- 
minated at that point.) This enrollment effort was much 
more successful than those experienced by other 

HMOs under Medicare contracts, which usually 
resulted in enrollment of only a few hundred benefici- 
aries in the first year. 

Fallon's second year open enrollment period lasted 
from September 15 through January 10, 1981. Approxi- 
mately 23,000 non-group Blue Cross/Blue Shield 
"Medex" subscribers received an application and 
brochure. Prior to this mailing, 900 beneficiaries 
on a waiting list were sent applications. The Fallon 
Clinic held five more well-attended open houses and 
gave several presentations for senior citizens groups. 
Newspaper advertising with return coupons was 
repeated, and marketing representatives contacted 
employer-retiree groups. A 20 percent penetration rate 
was achieved in a large industrial company with over 
500 retirees. Since the program began, approximately 
150 members have disenrolled. (This figure includes 57 
deaths.) At the close of open enrollment, 1,700 more 
Medicare beneficiaries were enrolled. The marketing 
projection for Medicare enrollment in the first quarter 
of 1981 was achieved, bringing Fallon's Medicare 
penetration in its service area to nearly 10 percent. 

The generous benefit package probably explains much 
of Fallon's enrollment success. Under the project, 
Fallon's Adjusted Community Rate was low enough, 
compared to area fee-for-service costs, that the plan 
could offer enrollees several benefits not covered by 
Medicare, including eyeglasses, preventive services, 
reduced coinsurance and deductible expenses, and 
reduced drug costs. Fallon's marketing techniques may 
also explain its success. 

MARSHFIELD: Marketing activities at Marshfield began 
in April 1980. Like Fallon, Marshfield concentrated its 
efforts on local meetings with elderly residents and 
senior citizen groups. In the Marshfield area, 29 
meetings were held in town halls, armories, and other 
public meeting places, and they were announced in 
advance in local newspapers and on the radio. The 
meetings were generally well attended and resulted in 
over 2,700 enrollments in the first three months. 
Approximately 3,550 people attended these meetings, 
indicating that over three-quarters of the people who 
attended the meetings actually joined the plan. 

In addition, Marshfield set up a special enrollment desk 
in the Marshfield Clinic, and of the plan's first 6,200 
enrollments, approximately 2,600 enrolled at the clinic. 
Marshfield speculates that this group may represent an 
adverse selection because people who enroll at the 
Clinic are users of services, almost by definition. 
Marshfield's total enrollment of 6,200 represents 
approximately one-third of all eligible Medicare benefi- 
ciaries in the service area. As of September 1980, 10 renal 
beneficiaries had also enrolled in the plan. Marshfield 
is holding continuous open enrollment. 

Unlike Fallon and Kaiser, Marshfield was not able to 
offer a particularly generous benefit package under its 
capitation rate. Its initial success was probably a func- 
tion of the popularity of the plan among the non- 
Medicare population. (The plan has approximately a 40 
percent penetration rate in the non-Medicare commun- 
ity.) Moreover, since the same providers render most 


fee-for-service as well as prepaid care in the area, 
people did not have to breal< existing provider relation- 
ships to join the HMO. Marshfield's second marketing 
campaign, which began in October, was not as suc- 
cessful as the first. The plan's goal was to enroll 12,000 
Medicare beneficiaries by March 1981. As of the end of 
February, 7,929 new members had joined. The plan 
mailed literature and held public meetings, but news- 
paper advertising was minimal. The demonstration 
attempted to attract more institutionalized enrollees 
through a special mailing to responsible family 
members or guardians. From 38 mailings, six persons 

KAISER: Kaiser's campaign featured TV ads, a first for 
Kaiser. There was some initial reluctance on Kaiser's 
part to run the ad because of a sensitive relationship 
with the rest of the medical community. Kaiser also 
sent letters to its Group Practice Prepayment Plan 
enrollees explaining the program and suggesting they 
share the information with eligible friends and relatives. 
Kaiser began its marketing efforts in late May 1980 and 
reached its first year enrollment projection of approxi- 
mately 6,330 by mid-summer. 

The reasons for Kaiser's enrollment success may be 
similar to Fallon's. Its ACR was low enough to enable it 
to eliminate deductible and coinsurance for enrollees, 
and to provide routine physical exams, exams for eye- 
glasses, most immunizations, and full coverage for pre- 
scribed home health care and non-psychiatric out- 
patient mental health services. In addition. Kaiser has 
an advantage in that it is a large, well-known 

As part of its first enrollment period. Kaiser conducted 
a study to determine which additional benefits were most 
attractive to potential enrollees. All beneficiaries were 
randomly assigned to one of two groups: 

Group I 

Medicare-coordinated (M-Plan) coverage without 
payment of dues 

Group II 
Choice of one of four options listed below: 

• Option A was the same as Group I above— no 

M-Plan dues (Options B-D required higher dues 
than Option A; however, they provided additional 
benefits not included in the M-Plan); 

• Option B provided prescribed drugs at $1 per 

prescription, prepaid eyeglasses, and hearing aids 
($6 premium); 

• Option provided dental care and dentures ($9.81 

premium); and 

• Option D provided prescribed drugs at $1 per pre- 

scription, prepaid eyeglasses, hearing aids, dental 
care, and dentures ($15.81). 

After two months of applications, enrollment in the 
various options was as follows: 

Benefit Option 
Medicare Plus (no premium) 

Percent Enrolled 

Medicare Plus and drugs/vision/hearing 
aid ($6) 

Medicare Plus and dental ($9.81) 

Medicare Plus and drugs/vision/hearing 
aid/dental ($15.81) 




Subsequently, all beneficiaries will be offered the choices 
available to Group II. 

Conversions and Disabled Beneficiaries 

In each demonstration, a few enrollees represent con- 
versions, that is, people who are members of the HMO 
under existing employer group contracts, and turn 65 
or become disabled. These people are easy to identify 
and contact because they are already HMO members. 
So far, they represent only a handful of people under 
the Fallon and Marshfield contracts, but they will grow 
in number during the demonstration. Kaiser has a large 
number of conversions from its existing GPPP contract 
with Medicare, which covers Part B services only. The 
number of GPPP conversions was limited to 1,500 out of 
the first 4,500 enrollees. 

An interesting finding from the Fallon and Marshfield 
demonstrations is that almost all the enrollees are 
aged, with very few disabled beneficiaries represented. 
This is not surprising, given the emphasis on the aged 
in the marketing materials and various benefits to the 
elderly of joining. Fallon actually refers to its program 
as the "Senior Plan." Although the marketing efforts 
have obviously been very persuasive with the elderly 
population, the results suggest that more could be 
done in targeting some of the marketing efforts to the 
disabled, who compose close to 10 percent of the 
Medicare population. 


HMOs may decide to pursue a Medicaid contract for 
several reasons. Some see it as a legitimate means to 
expand membership and increase revenues. Some view 
a Medicaid contract as a social responsibility. For 
some, all of these considerations are important. How- 
ever, while 30 States have HMOs within their bound- 
aries, only 17 contract with them. Obstacles to both the 
State and the HMOs to contracting to serve Medicaid 
recipients have been identified in this paper and else- 
where. "^ States are especially reluctant to inadvertently 

43. Morris, S., "Obstacles to Expanded Prepayment under 
Medicaid HMO Contracts" June 1980. Internal HCFA 


replicate the California experience. To avoid doing so 
they see a need to closely monitor HMO marketing 
practices, but cannot or do not wish to divert current 
staff or hire qualified persons to do this job. In States 
which do have the desire and capacity to manage HfvIO 
contracts, many HMOs simply do not feel that any 
advantages in contracting outweigh the difficulties in 
obtaining qualification, establishing capitation rates, 
assuming risk, fulfilling reporting requirements, and 
providing certain benefits which are part of the State 
plan but may place an untenable burden on the HMO 
(for example, EPSDT screening or dental care). 

Perhaps the most significant deterrent that an HMO 
considers before contracting with the State is how to 
market itself to the Medicaid population. The need for 
effective, inexpensive Medicaid marketing is important 
to HMOs for two reasons. First, although a core of 
enrollees usually retains Medicaid eligibility for appre- 
ciable periods of time, many others do not. Besides 
having to continually educate new people about using 
the system, the HMO must be able to attract these new 
enrollees on an ongoing basis. 

Second, and perhaps more importantly, most States' 
HMO rate-setting procedures do not provide sufficient 
funds for effective marketing. If a State sets HMO rates 
on a percentage of fee-for-service, it may not recognize 
that fee-for-service payments do not include a 
marketing component. If a State sets its rates 
actuarially, it may not be willing to pay more than it 
costs the HMO to market to the private sector. Such a 
position fails to consider that private sector marketing 
is done largely on a group basis— hence economies of 
scale are probably achieved. HGFA is currently explor- 
ing a regulatory change to allow cost contracts 
between States and qualified HMOs. If these regula- 
tions are eventually approved, they may provide an 
avenue for recognizing the full costs to HMOs of 
marketing to the Medicaid population. 

The problems in designing a Medicaid marketing 
strategy begin with identifying the target population. 
Title XIX confidentiality provisions prohibit the release 
of names. This had led a number of States, including 
California, to permit door-to-door canvassing in neigh- 
borhoods known to have large numbers of residents 
receiving welfare payments. Although expensive, door- 
to-door marketing can be effective in producing enroll- 
ments. Some have suggested that this is because the 
poor have traditionally responded best to peddling as a 
sales technique."" However, door-to-door marketing is 
generally viewed as too conducive to unethical pres- 
sure on the recipient to enroll. In California, the oppor- 
tunity for this was even greater in the early years of the 
program since the PHP salespeople were permitted to 
ifinalize enrollments in recipients' homes. Today, in 
those States where door-to-door presentations are still 
made, enrollments are completed elsewhere to ensure 

that the prospective enrollee fully understands the 
choice he or she is making. 

PHRED Membership Studies 

Besides door-to-door, two other marketing alternatives 
are currently used to enroll Medicaid recipients: (1) 
State mailing of brochures to recipients explaining that 
they have a choice between an HMO in their area and 
seeking care where they can find a provider willing to 
serve them, and (2) various presentation methods in 
the welfare office. The PHRED project, which is com- 
pleting the Membership Studies component of its grant, 
has shed some light on the comparative costs and 
benefits of these alternative marketing strategies. Be- 
sides developing model grievance and disenrollment 
systems for the plans, the major thrust of Membership 
Studies was to demonstrate and evaluate alternative 
methods of identifying the marketplace and motivating 
enrollment. The Membership Studies demonstrations 
were based on the assumption that the joint eligibility 
process for welfare and Medicaid affords the best pos- 
sibility for replacing door-to-door solicitation as a cost- 
effective locale for enrolling the number of recipients 
necessary for viable contracts."^ PHRED tested five 
marketing methods in seven welfare offices in northern 
and southern California. The welfare office methods 
included (1) a printed brochure with no personal 
explanation, (2) a film, (3) a personal presentation by a 
county eligibility worker, (4) a similar presentation by a 
specially trained member of the PHRED staff, and (5) a 
personal presentation by an HMO sales representative. 
In addition, literature was mailed to all eligibles in the 
geographic areas served by the demonstration sites' 
welfare offices. The plans' current marketing 
approaches (door-to-door, member referrals, conver- 
sions) served as the control. 

PHRED's preliminary evaluation of the 5,913 choices 
made in welfare offices under the demonstration has 
resulted in the following initial findings:"® 

(1) Marketing in the welfare offices can yield a suffi- 
cient enrollment of Medicaid AFDC beneficiaries 
to permit elimination of door-to-door solicitation 
without adverse effect on the viability of the 
health plan contracts. State-wide enrollment 
would increase as a result of the full implementa- 
tion of the demonstrated approaches, although 
impact on individual contracts may vary. Out of 
5,913 choices, 951 (or 16 percent) were for an 

(2) All methods tested produce enrollments, but 
some methods are more uniformly productive 
than others. Results of some seem to reflect local 
conditions rather than performance of the 

44. Greenberg, B.S. and B. Dervin, 1979. Use of the Mass 
Media by the Urban Poor. New York: Praeger: Caplovltz, D. 
1963. The Poor Pay More. New York: The Free Press of 

45. Owen, E., "A Preliminary Report on the Marketing Demon- 
stration Results from California's Prepaid Health Research 
Evaluation and Demonstration (PHRED) Project," 26 
September 1980, Unpublished. 

46. Ibid. 


method itself. Table 1 indicates the results of the 
5,913 presentations: 

Choice Rates by Method 

Welfare Office 


Choice Rate 

for HMOs 

by Sites 

Written Materials 






Eligibility Worker 



State Staff 



HMO Representative 



Other Methods 



4% response indicating 
interest or enrolling 
directly in HMO 




Costs for marketing in the welfare office are less 
than the plans' current methods. Welfare office 
enrollments cost about $9 per person enrolled, 
whereas door-to-door marketing costs between 
$45 and $50 per person enrolled. 

Follow-up interviews of 1,200 persons who made 
choices in the welfare offices indicated that those 
who chose HMOs were twice as likely to score 
more than 50 percent correct answers about their 
options and choice than those who chose fee- 

Medicaid beneficiaries in the welfare offices are 
able to make informed choices despite the pres- 
sure for funds, food, or other immediate needs 
which bring them to apply for welfare. Of 5,913 
presentations, only 41 did not choose. 

(6) A survey of some of these beneficiaries indi- 
cated that 22 percent elected the HMO to get 
"better service." About 21 percent thought the 
HMO would be more convenient, 13 percent be- 
lieved the HMO to be "better than the fee-for- 
service card," and 11 percent cited previous 
experience. Nine percent had no previous source 
of care, and 16 percent had other reasons. Of 
those choosing fee-for-service, 31 percent indi- 
cated that they did not wish to sever established 
provider relationships, 11.5 percent cited con- 
venience, 9 percent wanted freedom of choice, 
7.5 percent noted familiarity with the fee-for- 
service system, and 41 percent had other 

AB 1693, California's legislative response to the PHP 
scandals, authorized the Director of the Department of 
Health Services to prohibit door-to-door marketing 
after January 1, 1979, if suitable methods had been 
found to replace it. The law also enabled PHRED to 
test and evaluate new marketing techniques. The 
Department of Health Services is presently writing the 
regulations to abolish door-to-door marketing. Based 

on the experience of Membership Studies, the State 
plans to offer dual choice to Medi-Cal eligibles in the 
county welfare offices as part of the welfare determina- 
tion process. Presentations will be made by State-paid 
county employees who will be trained and closely 
monitored by the State. 

Although the data suggest the adoption of presenta- 
tions by HMO representatives, a number of drawbacks 
to that method limit its desirability in California, and 
perhaps elsewhere. These drawbacks include:"' 

• cost to plans 

• residual negative attitudes of welfare workers 
toward health plans 

• opposition by California Medical Association 

• limited space in welfare offices for more than one 
or two HMO representatives 

• the insignificant penetration difference resulting 
from presentations by State versus HMO 

There are two other legal factors which probably make 
the HMO representative method impractical. First, Title 
XIX confidentiality provisions would require that the 
HMO enroller not know the beneficiary's identity until 
the beneficiary chooses to reveal it. Second, no pro- 
vider can be accorded special access to patients. 
Therefore, if a welfare office permits an HMO to market 
on its premises, it may have to provide the same 
access to other providers. 

California is also anticipating the passage of a bill 
which calls for a mandatory choice by Medi-Cal 
eligibles during the welfare determination process. 
Enforcement of the mandatory choice, however, has 
generated controversy. At first, the State proposed to 
consider non-choosers as "entitled, but not covered," 
as a private employer would regard someone who had 
not selected a health benefits plan. The counties 
opposed this on the grounds that non-choosers would 
present themselves at the county hospitals with no 
evidence of coverage. The California Medical Associa- 
tion opposed the issuance of a special card to non- 
choosers in favor of the fee-for-service card. It now 
appears that non-choosers will get the latter card, but 
the State believes that the intent of the law— to place 
responsibility for choosing on the beneficiary— will still 
be maintained. 

Massachusetts Case Management 

Regardless of the relative effectiveness of the person 
or the medium used to encourage a Medicaid recipient 
to enroll in an HMO, there are disincentives to 
choosing an HMO which might be overcome only 
through legislative or regulatory change or, on a 
limited basis, through demonstrations. In the private 
sector, Luft has shown that the majority of HMO 
members enrolled for economic reasons. In effect, the 
HMO savings is being passed on to the enrollee in 
reduced premium payments, often accompanied by 

47. Conversation with Elizabeth Owen, PHRED Project 
Director, July 1980. 


additional benefits. Medicaid enrollees, however, can- 
not presently share in the savings realized by their 
joining the HMO or receive cash payments v\/hich 
would compensate them for limiting their care to one 
cost-effective provider, nor can an HMO offer addi- 
tional benefits not covered under the State plan. 

Preliminary evaluation results from the Massachusetts 
Case Management Grant, which offers a monthly $7 
incentive payment to AFDC recipient families who 
agree to enroll with a provider, indicate that shared 
savings provide an effective stimulus to enroll in a sys- 
tem which restricts access. Case Management was 
implemented in July 1979 at four fee-for-service pro- 
vider sites (two hospital-based and two community 
health centers). These providers share savings with the 
State between the annual costs of case-managed and 
non-case-managed care in the same region. Prelimi- 
nary cost and utilization analyses for three of the sites 
showed an $81,000 net savings at a community health 
center and hospital-based clinic and no savings at 
another hospital-based site. In October 1980, the State 
signed contracts with two new sites (one health center 
and one group practice) which have adopted risk- 
sharing reimbursement methodologies. As of that time, 
the project had enrolled 1,640 families (36 percent of 
the target set for April 1981). Increased marketing staff 
at the new risk-sharing sites, along with radio adver- 
tising, have been credited with more successful enroll- 
ment results at these sites. 

Guaranteed access to care is the chief selling point 
with which HMOs currently motivate Medicaid recipi- 
ents to enroll in their plans. Experience has shown that 
in most places, this simply is not enough to make a 
difference. ORDS is implementing demonstrations in 
Hawaii and California to determine whether guaranteed 
eligibility, as well as cash payments, can serve as addi- 
tional inducements for Medicaid eligibles to enroll in 

Multnomah County (Oregon) Project Health 

An alternative approach to marketing, the broker con- 
cept, was introduced by Multnomah County Project 
Health. The goal of the project is to permit the State of 
Oregon to test the effectiveness of a multiple choice 
delivery system for the medically needy, ultimately 
leading to passage of legislation of a state-wide 
medically needy program. The project has also func- 
tioned as a laboratory for testing the response of low 
income families to trade-off decisions, where the health 
plan of choice may mean more copayments and/or 
higher premiums. 

As Project Health evolved, the county developed a sys- 
tem of contracts with the private provider community 
involving (1) full-risk capitation contracts with prepaid 
health plans and (2) fee-for-service contracts with 
individual providers. HCFA has supported the project 
since January 1976 by granting waivers to the State's 
medical assistance plan. Since 1976, Project Health has 
served an average of 2,000 medically needy benefici- 
aries. Seventy-five percent of the medically needy 
population is currently served by prepaid contracts. 

At the time of enrollment, those beneficiaries not auto- 
matically assigned to the fee-for-service program 
because they are institutionalized are given a choice 
among the prepaid plans. If the beneficiary chooses 
the lowest cost prepaid plan, the premium for that plan 
is paid entirely by Project Health. If the beneficiary 
elects a higher cost plan, then s/he shares in the 
premium cost. The amount of premium share is 
dependent upon, but not equal to, the difference 
between the chosen plan's premiums and those of the 
lowest cost plan. Prepaid plans may also use 
copayment and deductibles as beneficiary cost-sharing 
methods. All members of Project Health are given a 
comprehensive scope of benefits, regardless of the 
plan selected. Table 2 shows that beneficiaries will 
elect to enroll in higher cost plans despite the premium 

Comparison of Enrollment in Project Health Prepaid Plans' 

Project Health 








Average Quarterly 


Client Monthly 


of Total 

Enrollment Fee for 


Number of Famil 



Family of 









$ 98 




























*Data taken from Third-Year Evaluation of the Medicaid 
Demonstration Project in Multnomah County, Touche Ross and 
Company, September 30, 1979. 


The project has shown that medically needy benefici- 
aries are willing and able to share in the cost of prepaid 
plan premiums. This willingness is particularly evident 
when cost-sharing is necessary to maintain a pre- 
existing provider-patient relationship. 

Quality of Care 

Current Procedures 

Section 1301(c) of the HMO Act of 1973 provides 
authority to establish the quality assurance require- 
ments HMOs must meet to become Federally-qualified 
and remain in compliance. Since Medicare and 
Medicaid contract almost exclusively with qualified 
HMOs, these statutory and regulatory requirements 
affect Federal beneficiaries as well as employed 
persons enrolled in HMOs. 

Section 1301(c)(8) requires that each qualified HMO: 

"have organizational arrangements, established in 
accordance with regulations of the Secretary, for an 
ongoing quality assurance program for its health 
services which program (a) stresses health outcomes, 
and (b) provides review by physicians and other 
health professionals of the process followed in the 
provision of health services:" 

Section 1301(c)(11) further requires that each qualified 

"provide, in accordance with regulations of the 
Secretary (including safeguards concerning the con- 
fidentiality of the doctor-patient relationship), an 
effective procedure for developing, compiling, evalu- 
ating, and reporting to the Secretary, statistics and 
other information (which the Secretary shall publish 
and disseminate on an annual basis and which the 
health maintenance organization shall disclose, in a 
manner acceptable to the Secretary, to its members 
and the general public) relating to (a) the cost of its 
operations, (b) the patterns of utilization of its ser- 
vices, (c) the availability, accessibility, and accepta- 
bility of its services (d) to the extent practical, 
developments in the health status of its members, 
and (e) such other matters the Secretary may 

The Office of Health Maintenance Organizations 
(OHMO) is currently implementing a quality assurance 
strategy for HMOs based on requirements for an 
internal and external program.-- The broad require- 
ments for an internal quality assurance program stipu- 
late that: 

• The program must be supervised by a physician in 
the HMO and involve a broad spectrum of health 
professionals in the HMO. 

• The responsibility for review cannot be delegated 
to an outside party. 

• Current activities of the HMO, such as personnel 
selection and laboratory controls, may be identified 

48. "Quality Assurance Strategy for HMOs." OHMO. USPHS, 
DHHS. Rockville, Maryland, September 1. 1979. 

as part of the HMO's overall quality assurance 
• The program must provide for a systematic and 
documented approach involving both inpatient and 
outpatient settings. 

OHMO suggests Medical Care Evaluations (MCEs) as a 
flexible way to meet these requirements which can be 
easily integrated with HMOs' other quality activities. 
MCEs offer a single review approach, regardless of the 
source of the premium payment. Moreover, the cost of 
MCEs can be recognized in the capitation as an over- 
head or administrative expense. OHMO recommends 
that HMOs' quality assurance programs not be limited 
to physician-related health care, but include care pro- 
vided by all professionals in the HMO. Topics can be 
selected in any area where problems may exist, but 
should be relevant to patient well-being. The responsi- 
bility for selecting topics for MCEs and conducting 
them should rest with "a discrete committee within the 
organization with clearly defined lines of communica- 
tion to those responsible for implementing changes." 
While OHMO sets no numerical requirement, it does 
indicate that each HMO should implement a program 
which systematically examines all major areas of care 
and prioritizes problems for study and corrective 

Although a number of well-established HMOs have 
quality assurance programs which could easily meet 
these guidelines. OHMO anticipated that developing 
HMOs would require technical assistance in imple- 
menting an internal program and that they may also 
require 12 to 18 months' experience with their internal 
programs before an external assessment is conducted, 
In November 1979, OHMO awarded a 367,000 technical 
assistance contract to InterOual, Inc., Chicago. The 
firm has since provided over 20 on-site consultations to 
HMOs which requested their assistance and is prepar- 
ing a technical manual on how to design an HMO 
quality assurance program. 

The second component of OHMO's quality assurance 
strategy — external review— will ensure that qualified 
HMOs can assure quality of care. The external review 
strategy calls for delegation of responsibility to an 
independent organization. OHMO is negotiating a con- 
tract with the National Committee for Quality Assur- 
ance, Inc. to develop review procedures and standards. 
Consulting physicians trained by the National Commit- 
tee will conduct on-site HMO reviews. Each HMO will 
be externally assessed every three years: however, if 
evidence of a serious interim problem surfaces, a "for- 
cause" assessment would be conducted upon referral 
by OHMO to the National Committee. 

The physicians who review particular HMOs will be 
from similar type HMOs (that is, group, staff, or IPA). 
The National Committee will report to OHMO on the 
status of ongoing quality assurance efforts in HMOs 
and investigate "for cause" under a detailed protocol 
for ensuring confidentiality. Approximately 12 routine 
assessments are scheduled for fiscal year 1981. The 
Committee will also attempt to define and identify 


measurement techniques for "availability, accessibility, 
and acceptability" of services. 

HMOs and PSROs 

Section 1155(e)(1) of the Social Security Act authorizes 
Professional Standards Review Organizations (PSROs) 
to use the services and accept the findings of the 
review committee of a ". . . health care facility or 
organization." In the past, PSROs have delegated con- 
current review and MCE functions to some hospitals, 
some of which are owned by HMOs. However, a recent 
HHS General Counsel decision indicates thatthe law also 
authorizes PSROs to delegate directly to HMOs the 
responsibility for review of their members, including 
those HMOs which do not own the hospitals to which 
their patients are sent. HCFA is writing a new chapter 
of the PSRO Program Manual which will set forjth the 
criteria for delegation, as well as the agreements that 
should be signed by the PSROs, HMOs, and hospitals 

For an HMO to obtain responsibility to review its 
Federal enrollees, it must request permission from the 
PSRO. Unless waivers are granted, a delegated HMO 
must use PSRO norms and criteria for reviewing 
Federal enrollees. In delegating review to an HMO, 
PSROs should recognize that, while the fee-for-service 
system review focuses on over-utilization, the review of 
HMO patients should focus on potential under- 

The rationale for allowing HMOs to conduct their own 
review is partly to avoid duplication of effort in the 
review of Federal beneficiaries. Delegation should also 
eliminate the potential for a PSRO to deny a hospital 
bill for a Medicare beneficiary enrolled in an HMO with 
a risk contract. In such a situation, the unpaid hospital 
may turn to the beneficiary for reimbursement, thus 
abrogating the agreement between the HMO and the 
beneficiary that the HMO will pay for all in-plan 
services it orders. 

It is not clear how willing PSROs will be to delegate 
review. The PSRO responsible for the review of 
patients enrolled in the Marshfield Medical Foundation 
(one of HCFA's demonstration sites) has thus far 
opposed Marshfield's requests for delegation. The 
Central Massachusetts PSRO which reviews the 
medical necessity of care provided to Fallon Medicare 
enrollees also opposes delegation to Fallon at this 
time. The circumstances here, however, raise a poten- 
tial conflict of interest, since the membership of this 
particular PSRO comprises many of the same physi- 
cians who have formed an IPA known as the Central 
Massachusetts Health Care Foundation. ••' Also 
problematic is whether many HMOs will want PSRO 
review responsibility, since no provision has been 
made for the PSRO to reimburse the HMO for its 
services. HMOs may view this policy as inequitable, 
given that hospitals with delegated review functions are 

49. Lavln, J. H., "Doctor Surplus: Close-Up of a Town That's 
Feeling the Crunch," Medical Economics, September 29, 
1980, pp. 69-80. 

Medicare and Medicaid 

The Medicare program relies on the internal quality 
assurance systems of the HMOs with which it contracts 
and the oversight responsibility of the OHMO. HCFA 
conducts no independent review of the quality assur- 
ance capability of the HMOs with Medicare contracts. 
On the other hand, the Medicaid program does require 
States contracting with HMOs to assume direct respon- 
sibility for quality of care. Regulation 42 CFR 431.513 
requires the State agency to "establish a system of 
periodic medical audits to insure that the HMO pro- 
vides quality and accessible care to enrolled recipi- 
ents." The regulations further require that the State (1) 
conduct an audit at least annually at each HMO, (2) 
collect management data for medical audit personnel, 
and (3) ensure that the data include reasons for enroll- 
ment and termination and use of services. 

One of HCFA's long-term policy objectives, to the 
fullest extent practical, has been to coordinate the 
operating procedures for the Medicare and Medicaid 
programs. Quality assurance requirements for Medicare 
and Medicaid HMO enrollees might be one area where a 
more uniform policy would be appropriate. Since nearly 
all Federal beneficaries are receiving is already prepaid 
health care are enrolled in qualified HMOs, the quality of 
care these beneficiaries are receiving is already subjectto 
OHMO quality assurance, requirements. The Medicaid 
program, in effect, calls upon States to replicate the 
effort of OHMO. The quality assurance requirements of 
the Medicaid program may even be inhibiting the 
development of contracts between States and HMOs. 
States may be less inclined to pursue Medicaid 
contracts if they are required to set up new quality 
assurance programs. Similarly, HMOs already meeting 
OHMO quality assurance requirements may view one 
more set of standards and data requirements as not 
worth the cost and effort. 

The PIHRED Quaiity Assurance Demonstration 

In addition to rate-setting and enrollment studies, the 
third major component of the PHRED project was to 
demonstrate and evaluate two alternative quality assur- 
ance strategies: an encounter data system (the monitor 
aooroach) and a selective data acquisition approach 
based on traditional medical audit review techniques. 
Both systems apply computerized criteria logic to age, 
sex, diagnosis, and service data and report the per- 
formance of the HMOs in passing or failing the criteria for 
the services examined. The first task in implementing the 
demonstration was to identify the criteria of performance 
to be applied to both information systems. Over a period 
of several months, a panel of 12 experts organized a 
set of 332 criteria grouped into 20 sections according 
to ICDA-8 diagnostic codes. The criteria set was 
reviewed by numerous local professional organizations 
and individuals. Three California HMOs with Medicaid 
contracts agreed to participate in the demonstration. 
The two data acquisition approaches each use a differ- 
ent subset of these 332 criteria. 

PHRED project staff designed and implemented the 
monitor approach. In doing so, PHRED adapted the 
claims-based Physicians Ambulatory Care Evaluation 


(PACE), which is an approved Medicaid Management 
Information System operated for the State of Utah by 
the Utah Professional Review Organization. The key 
characteristic of the monitor method is that it uses 
information on all patient encounters submitted by the 

To operate the system, PHRED developed computer 
modules to collect and edit encounter and Medicaid 
eligibility data, to screen 250 criteria against the 
encounter data, and to generate potential exceptions to 
quality of care by the patient and/or provider. These 
systems can maintain and update patient profiles of 
care rendered and produce reports reflecting HMO and 
physician performance levels. 

The selective approach was developed and tested by 
the Kaiser Research Center under contract with the 
State of California. The system uses a set of criteria 
programmed into a mini-computer. The program's 
branching logic instructs nurse abstractors to look for 
specific conditions in a patient's chart that will satisfy 
the question posed. The system reports "passes" and 
"failures" to the criteria and outlines performance rates 
for the criteria in each plan. Kaiser is preparing a 
report on the selective data acquisition system which 
will describe the history and results of the mini- 
computer application of 125 criteria derived from the 
PHRED criteria set. 

As of December 1980, all data had been collected from 
the demonstration sites. SysteMetrics, Inc. is evaluating 
the two data acquisition systems to identify the pre- 
ferred approach for California. The criteria against 
which the two systems are being evaluated include: cost, 
completeness, data validity, and acceptability to the 

The evaluation is scheduled to be completed in spring 
of 1981. Each system is expected to have benefits and 
drawbacks. The selective system is probably less 
expensive and easier to implement, since it requires no 
changes in HMO operations. Implementation of the 
monitor system would require more front-end costs as 
well as modifications to current HMO reporting sys- 
tems. However, the information generated by the 
monitor system, once in place, could facilitate rate- 
setting and performance evaluation in areas outside of 
medical quality. California has accepted a budget pro- 
posal prepared by PHRED staff to implement the monitor 
system into the State agency's review activities. 

Program Policy Considerations 

All available information indicates that Medicare enroll- 
ment will not accelerate appreciably unless major 
reforms in reimbursement are implemented. However, 
experience in implementing demonstration projects has 
highlighted the need for greater flexibility in the 
policies and procedures which HMOs will be required 
to follow in interacting with HCFA, intermediaries, 
carriers, providers, and beneficiaries under a full risk 
reimbursement process. These procedures, problems, 
and alternatives are discussed below. 

Claims Processing 

The current system for reimbursing HMOs under 
Medicare draws heavily on procedures developed for 
fee-for-service reimbursement. The major reason is 
that both cost and risk formulas authorized by Section 
1876 involve reimbursement on the basis of the cost of 
services rendered. Neither method is prospective. Con- 
sequently, the administrative procedures currently in 
place are not always appropriate for reimbursement 
methods not tied to specific services rendered, as in 
the case of the demonstration projects. 

An example is the set of procedures an HMO must 
follow in paying providers for Part A services. Under 
current Section 1876 risk contracting, an HMO may 
elect to pay all Part A claims except those for dialysis 
and related services. This option (referred to in the 
HMO Operating Instructions as "Option C") is the one 
selected by each of the demonstration projects. Each 
demonstration HMO is requesting the providers it deals 
with to submit claims on its own forms for processing 
by the HMO. However, under Section 1876 procedures, 
HMOs are also required to obtain the Medicare in- 
patient billing form (HCFA-1453) from each provider, 
and if payment is made for the service, to forward the 
bill to the appropriate intermediary for transmittal to 
Baltimore. A HCFA-1453 is required to permit updating 
of the beneficiary's master record with utilization infor- 
mation, so that individual spell of illness and lifetime 
benefit limitations can be tracked accurately. These 
utilization indicators become important if the benefici- 
ary disenrolls from the HMO. 

HMOs and providers have both objected to this 
procedure because it creates additional paperwork for 
providers, who must often prepare two billing forms, 
and introduces some chance for duplicate payment 
and administrative mix-ups. A more appropriate proce- 
dure may be to allow the HMO to require its own 
billing forms from in-area providers and not to dictate 
that HCFA-1453 forms also be submitted. Under this 
arrangement there would be no transmittal of HCFA- 
1453s to Baltimore or updating of the beneficiary's 
record as long as s/he was enrolled. A similar proce- 
dure could be followed for out-of-area provider claims, 
that is, they would not be returned to the intermediary 
for transmittal to Baltimore to update the beneficiary's 
record. In the majority of cases, the beneficiary would 
not be disadvantaged if his/her master record were 
"frozen." Rather, it would benefit many. It is probably 
not worth the administrative cost to track all these bills 
under a prepaid arrangement. 

Another potential administrative problem is the 
handling of claims for out-of-plan services. Under 
claims processing Option C, these claims (for both 
Part A and Part B services) are processed and paid by 
the HMO. If a claim is sent by the provider or 
beneficiary directly to the intermediary (or carrier), the 
intermediary will receive a message in the query reply 
indicating that the beneficiary is enrolled in an Option 
C risk-basis HMO and that the claim should be sent to 
the appropriate HMO. The problem is that there is 
nothing to prevent the intermediary or carrier from 
actually making payments on such a claim, rather than 


forwarding it to the HMO, as instructed. If the inter- 
mediary inappropriately makes payment for an out-of- 
plan service, there is no mechanism for withholding 
some reimbursement from the HMO for that service 
under a prepaid arrangement. In addition, if duplicate 
claims are sent to the intermediary and the HMO by 
the pi'ovider, duplicate payment could result for the 
same service. 

Coverage of Non-Medicare Services 

HMOs generally feel hampered by the restrictions 
imposed by the current Medicare benefit package. If an 
HMO believes it can provide the current services 
without imposing current restrictions, and still remain 
within its rate, it should be permitted to do so. For 
example, SNF services are normally covered only if the 
stay is preceeded by a three day covered hospital stay. 
The primary purpose for this requirement is to provide 
some guarantee that lower levels of care in SNFs are 
not covered under fee-for-service reimbursement. 
Under capitation reimbursement, however, such safe- 
guards are not necessary, and it could be left up to the 
HMO to determine which stays require a covered level 
of care, regardless of prior hospital stay. All the 
demonstrations have requested a waiver of the three 
day requirement, and the waiver has been approved in 
each case. Additional examples could include coverage 
of physician extender services and provision of home 
services by nurses. 

High and Low Option 

Section 1876 requires that HMOs offer the basic Medi- 
care benefit package to their Medicare enrollees and 
only offer additional benefits on an optional basis for a 
supplemental premium. The demonstration HMOs have 
requested waivers to permit them to offer one compre- 
hensive package only, which includes benefits normally 
not covered by Medicare. The beneficiary pays for 
some of these extra services through his/her premium, 
and the remainder are covered through the capitation 
rate. The HMOs believe that the requirement to offer a 
"low option" is administratively burdensome. The basic 
Medicare benefit package is not consistent with normal 
HMO comprehensive benefits, and HMOs are not 
generally prepared to administer them. Beneficiary 
interest in low option plans appears very minimal 
based on the GHC experience and that of HMOs with 
cost contracts. 

Premium for Low Option 

Current regulations prohibit the HMO from charging a 
premium for low option which is higher than the actu- 
arial equivalent of coinsurance and deductible. Under 
cost reimbursement, the premium is determined pro- 
spectively and subsequently adjusted to reflect actual 
utilization. For full risk reimbursement, the premium is 
not adjusted retrospectively. Often the prospective 
premium is considerably lower than the comparable 
premium charged by private insurers for similar 
benefits. HMOs argue that the premium should not be 
set to the actuarial equivalent, but to the rate required 
by the HMO to generate required revenues. A possible 

compromise is to require the HMO to inform the bene- 
ficiary of the actuarial equivalent of coinsurance and 
deductible, but to permit the HMO to charge a higher 
rate, as long as it is lower than the lowest private 
insurer rate. 

institutionalized Beneficiaries 

An additional issue with some HMOs has been the 
requirement that hospitalized or institutionalized bene- 
ficiaries must be allowed to enroll. This is counter to 
some HMOs' policies. Because the AAPCC and the 
reimbursement rates can be adjusted appropriately, it 
may be advisable to allow HMOs not to enroll institu- 
tionalized beneficiaries if this is their practice with their 
non-Medicare enrollees. On the other hand, if HMOs 
wish to play a significant role in providing health care 
for an increasingly aged American population, long- 
range interests of both HMOs and beneficiaries may be 
better served by enrolling the institutionalized. 

The issue of hospitalized beneficiaries is more 
complex, since it raises the question of favorable selec- 
tion. HMOs argue that they should not be required to 
assume responsibility for a Medicare beneficiary in the 
middle of a hospital stay, since this can disrupt the 
current beneficiary-physician relationship. In the 
private sector, the HMO would not assume responsi- 
bility until the patient was discharged. 

Renal Beneficiaries— Coinsurance and Deductible 

Under each demonstration, a single enrollee premium 
is calculated based on the average actuarial equivalent 
value of deductible and coinsurance expenses, plus the 
value of non-Medicare covered benefits not covered by 
the capitation rate. All enrollees are charged the same 
rate, even though the capitation rate may differ for dif- 
ferent age and sex groupings of enrollees. If renal 
beneficiaries are enrolled, as is the case with the 
Marshfield project, the renal enrollee will pay much 
less than the actuarial equivalent of renal-related 
deductible and coinsurance. (This fact could cause 
most renal beneficiaries in the area to enroll.) The 
result is that either the HMO or the other beneficiaries 
must subsidize this group. Otherwise renal enrollees 
would have to pay an extremely high premium. A 
similar problem arises even if the HMO does not enroll 
renal beneficiaries (as is the case with the other 
projects) because an aged or disabled beneficiary may 
develop renal disease. 

Renal Beneficiaries— Master Record Annotation 

HOFA master records of over-65 Medicare beneficiaries 
who develop end stage renal disease (ESRD) are often 
not annotated for several months after the beneficiary 
begins receiving routine dialysis. Thus, demonstrations 
such as Fallon, which are not at risk for renal 
beneficiaries, could face cash flow problems should a 
number of their enrollees develop permanent kidney 
failure. This is because cost reimbursement to the 
HMO for such enrollees can be initiated only after 
verifying that the beneficiary's master record is ESRD- 
annotated. Similarly, demonstrations such as Marsh- 
field, for which separate rates have been calculated for 


renal beneficiaries, would face perhaps months of 
waiting for record annotation to trigger the rate which 
incorporates the much higher renal costs. To address 
this problem in the demonstrations, the HCFA office 
which updates master records will receive special 
requests to annotate a beneficiary's record upon provi- 
sion of documentation that a particular enrollee has 
permanent kidney failure. While this system should be 
adequate for the few HMOs in the demonstrations, it 
may not be feasible for an expanded number of HMO 
Medicare contracts. 

Cost Reporting 

Under current cost and risk arrangements, HMOs are 
reimbursed on the basis of cost, with a consequent 
need for them to prepare quarterly and annual cost 
reports. Under a capitation arrangement, cost reports 
are not necessary because they are not used in deter- 
mining the next year's rates. Under the demonstrations, 
cost data will be collected for evaluation purposes only, 
and will fulfill no operational needs. 

HMO Medicare Manual 

The lack of a formal HMO manual has contributed 
to confusion on the part of intermediaries and HMOs. 
Intermediaries in particular have not demonstrated a 
knowledge of HMO claims processing and reimburse- 
ment procedures. 

Medicare Cards 

Several HMOs have expressed strong reservations 
about allowing beneficiaries to retain their regular 
Medicare cards, and would prefer to issue their own 
cards instead. These HMOs argue that a special card 
would curtail out-of-area usage by identifying a benefi- 
ciary as a lock-in HMO member to out-of-area pro- 
viders. Out-of-area usage has been a serious problem 
for GHC, resulting in numerous administrative prob- 
lems and dissatisfied beneficiaries. A special card 
would also serve as a reminder to beneficiaries of the 
lock-in provision. Problems with this possibility include 
the difficulty of persuading beneficiaries to relinquish 
their Medicare cards. (Even if they do, they can request 
that another card be issued.) Another problem is the 
current unfamiliarity on the part of many providers with 
anything except the normal Medicare card. 


Administrative procedures have not been as serious an 
issue under Medicaid because of the relative flexibility 
of the States in administering their programs. Nonethe- 
less, one problem which has been cited by a few 
HMOs currently holding Medicaid contracts is the 
impact of high disenrollment rates caused by loss of 
Medicaid eligibility. This increases the HMO's adminis- 
trative cost and often forces the HMO to retroactively 
seek reimbursement from other sources for services 
rendered to enrollees who were later demonstrated to 
be ineligible for Medicaid benefits. This problem can 
be particularly acute in those instances where the State 
is slow to notify the HMO of loss of eligibility. Unfor- 
tunately, HCFA does not have data on numbers or 

rates of disenrollment due to loss of eligibility in vari- 
ous HMOs or various States, but HMOs have frequently 
cited this as a significant problem. Demonstrations are 
being initiated in California and Hawaii to test the con- 
cept of guaranteed six month eligibility for Medicaid 
recipients who enroll in an HMO. A demonstration is 
also planned in Massachusetts in which eligibility for 
dental services only will be guaranteed to recipients 
who voluntarily enroll with participating prepaid dental 

Knowledge Gaps 

in the previous sections of this report, we have primar- 
ily addressed what has been learned thus far about 
Medicare and Medicaid experience with HMOs. Several 
important issues and questions remain, the answers to 
which should greatly facilitate growth of public sector 
contracting with HMOs. These issues are described 


What Is the Impact of full risk reimbursement? Many 
HMOs have claimed that if the Federal government 
would only pay them like the private employers, pros- 
pectively, with no adjustments based on actual costs, 
then more HMOs would contract to serve Medicare 
beneficiaries. However, these claims are made with 
little, if any, knowledge about the real impact of full 
risk reimbursement. While the Group Health Coopera- 
tive generated a remarkable savings in the first year of 
its Section 1876 risk contract, its means were contro- 
versial. GHC is an established HMO with its own hospi- 
tal; other HMOs may not be able to replicate its 
savings. Moreover, the GHC savings have decreased 
over the years, so that its costs have risen from 78 to 
90 percent of the AAPCC. Similarly, Fallon's second 
year adjusted community rate rose from 90 to 95 per- 
cent of the AAPCC. Although the HMOs participating 
in the capitation demonstrations do not constitute a 
representative sample of all HMOs, a relatively large 
number initially were not able to project large savings 
which could be turned into additional benefits. We will 
not have any more than anecdotal information about 
the impact of full risk reimbursement until the current 
demonstration projects have been evaluated, although 
we expect to report preliminary findings by mid-1982. 

How can Medicare beneficiaries best be encouraged to 
enroll in HMOs without a richer benefit pacltage than Is 
currently covered under Title XVIII? The demonstra- 
tions thus far have shown that additional benefits 
and/or reduced premiums are a strong incentive for 
Medicare beneficiaries to join an HMO. This finding 
essentially corroborates the observations of researchers 
studying employed HMO members that economic 
benefits are the chief attraction to joining an HMO. 
This incentive is recognized in the proposal to return 
to beneficiaries the difference between 95 percent of 
the AAPCC and the adjusted community rate in the 
form of additional benefits or reduced premiums. How- 
ever, what happens if an HMO with a Medicare contract 
cannot generate such a savings? Without the additional 
benefits/reduced premiums, it may be just as difficult 
to attract the Medicare beneficiary as it is under 
cost reimbursement. The problem may be especially 


serious for young HMOs that have not yet maximized 
efficiency and for HMOs whose AAPCC is low due to 
possible under-utilization or access barriers in the fee- 
for-service system (for example, Marshfield). While the 
Medicaid beneficiary may be motivated to join an HMO 
to achieve better access to providers, this is not 
generally a problem for Medicare beneficiaries. To the 
contrary, one expects Medicare beneficiaries to be 
reluctant to sever access to already established doctor- 
patient relationships. 

How many HMOs would be interested In a full risk 
reimbursement contract under Medicare? We do not 

know whether more than a handful of HMOs would 
enter into full risk contracts with HCFA, given the 
opportunity. However, it seems safe to hypothesize 
that, as in the private sector, if several HMOs can suc- 
cessfully demonstrate their viability under full risk, 
more would shortly follow. 


What is the most effective means of motivating Medi- 
caid recipients to enroll In HMOs? State of the art 
wisdom regarding alternative methods of stimulating 
Medicaid enrollment in HMOs offers the following 
marketing choices: (1) door-to-door, (2) mailings, (3) 
welfare office presentations, and (4) incentives such as 
cash payments and/or guaranteed eligibility combined 
with any of the preceding methods. The PHRED 
project's tentative findings argue in favor of welfare 
office presentations on the grounds that 16 percent 
penetration provides, across the board, sufficient 
enrollees to maintain viable health plan contracts. 
However, some HMOs will not be as well off if door-to- 
door solicitation is abolished, particularly those plans 
that were running aggressive door-to-door campaigns. 
On the other hand, PHRED claims that while those 
plans may suffer, other plans which previously did not 
contract because they did not want to market door-to- 
door will now sign contracts. Overall, HMO penetration 
will remain constant or increase. It will be important to 
look more closely at the impact on individual HMOs of 
abolishing door-to-door marketing, as well as to 
examine the analyses of the complete data set from 
Membership Studies, before positing firmer results. 
The value of offering guaranteed eligibility and cash 
payment as incentives to enrollment must be evaluated. 
A clear answer to this question awaits testing of incen- 
tives, as well as replication of the PHRED experience in 
other States. 

IVIedicare and IVIedicaid 

Do HMOs favorably select? The answer to this question 
has crucial significance to the future of HMOs and 
public payer participation. Proponents of both favor- 
able and unfavorable selection have valid arguments. 
This is clearly an area where future research is war- 
ranted in both the Medicare and Medicaid programs. 

if favorable selection occurs, Is a health status factor 
important to Incorporate into calculation of AAPCC? 

Given the complexity of devising a health status factor 
and collecting the data necessary to use it, policy- 
makers must ask whether health status is sufficiently 
influential, in and of itself, to warrant inclusion in the 
AAPCC. It may be that the AAPCC factors currently 
used (age, sex, welfare, and institutional status) 
account for most of the effect of health status. 

If health status should be Included In the AAPCC, 
how? Currently, no existing or planned Medicare data 
base contains an acceptable health status measure. 
Until the controversy over measuring quality of care is 
resolved, it is unlikely that there will be agreement 
about the proper means of measuring health status. 

Future Research 

Based on the above questions, the future research and 
demonstration agenda is likely to comprise the follow- 
ing activities: 

(1) Complete the current capitation demonstrations 
and evaluate them with particular emphasis on the 
selection issue, the impact of full risk reimburse- 
ment, and the effectiveness of alternative market- 
ing strategies. In assessing whether selection 
occurred under the demonstrations, it will be 
necessary to develop an acceptable measure of 
health status which may later be incorporated into 
prospective calculation of the AAPCC. 

(2) Conduct demonstrations to determine the impact 
of guaranteed eligibility and incentive cash pay- 
ment on Medicaid enrollment. 

(3) Compare utilization and cost experience of Medi- 
care enrollees enrolled under cost and risk con- 
tracts. The results of such studies may shed some 
light on whether the claims of HMOs of greater 
efficiency under full risk as opposed to cost are 

(4) Implement new Medicare risk-sharing demonstra- 
tions with organizations which do not fully satisfy 
the HMO definition, but have the capability of 
achieving some of the efficiencies of HMOs. 

(5) Conduct demonstration projects to test competi- 
tion models and compare how HMOs perform to 
other organizations providing a comprehensive 
health service package. 


The authors wish to express their appreciation to Barbara 
Cooper, Director, Office of Demonstrations and Evaluations, 
who provided many valuable comments on various drafts of 
this paper, and to Donna Eskow, who made suggestions 
which improved its style and literacy. Finally, we want to 
thank Lynn Betz and Iris Bridge, who flawlessly and 
uncomplainingly typed the many drafts of this report. 


Appendix A: Descriptions of l\/ledicare 
Capitation Demonstration Projects 

Fallon Community Health Plan 

The Fallon Community Health Plan is a qualified HMO 
jointly sponsored by the Fallon Clinic and Blue Cross 
of Massachusetts. As of September 1980, it served 
30,000 members, including 200 AFDC beneficiaries, in 
the Worcester area. Fallon is a one-group staff model 
HMO employing over 50 staff physicians. By the end of 
its demonstration contract in December 1982, Fallon 
hopes to achieve a 15 percent penetration in its service 
area, yielding a total enrollment of 6,000 Medicare 
beneficiaries with Part A and B, as well as Medicaid 
Old Age Assistance recipients. 

Marketing efforts to date have been targeted at group 
and non-group "Medex" subscribers, offering a dual 
choice between current coverage and that available 
through the Senior Plan. Senior Plan marketing efforts 
began on February 7, 1980, and within a month, the 
first year enrollment projection of 3,600 had been 
achieved. The marketing strategy focused on a "Health 
Fair" at the Fallon Clinic held prior to their enrollment 
campaign. The Health Fair offered presentations and 
talks about the plan, tours of the Fallon facilities, and 
booths dealing with a variety of topics generally related 
to the health and welfare of the elderly. Approximately 
4,000 people attended. Fallon also held five open 
houses and ran newspaper ads. The ads contained a 
card for interested readers to send in to receive a bro- 
chure and application. Fallon received over 3,000 
returned cards. In addition, Fallon obtained a mailing 
list of Medicare supplemental policyholders from Blue 
Cross of Massachusetts, which it used to disseminate 
literature in Worcester County. 

Fallon's second open enrollment campaign began on 
September 15 and ended on January 10, 1981. The 
plan enrolled 1,700 new members during that period. 
About 23,000 non-group "Medex" subscribers received 
a brochure mailing, along with 900 people on a waiting 
list. Fallon held five open houses, placed newspaper 
ads, and made presentations to employer-retiree 
groups. Fallon decided not to hold a health fair based 
on its marginal benefits relative to costs. 

Fallon's rates are based on an adjusted community rate 
(ACR). During the first year, its ACR was low enough, 
compared to area fee-for-service costs, to permit the 
coverage of several additional benefits at a $7.50 
monthly premium while the plan was reimbursed at 90 
percent of the Area Prevailing Cost. (The Area 
Prevailing Cost is the AAPCC where no adjustment is 
made for the demographic characteristics of HMO 
enrollees.) The additional benefits included: prescrip- 
tions with $1.00 copayment, eye examinations and one 
pair of eyeglasses, preventive services, and reduced 
coinsurance and deductible expenses. This generous 
benefit package probably explains the success of 
Fallon's enrollment efforts. Fallon offered beneficiaries 
the same benefit package at the same premium when 
new rates went into effect January 1981, even though 

the plan's reimbursement rate increased to 95 percent 
of the AAPCC. 

Marshfield Medical Foundation 

The Greater Marshfield Community Health Plan 
(GMCHP) is a non-Federally qualified HMO which was 
established in 1971. It is sponsored by the Marshfield 
Clinic (a 160-physician, multi-specialty, group prac- 
tice), St. Joseph's Hospital, Blue Cross of Wisconsin, 
and Surgical Care Blue Shield. The plan services over 
55,000 residents in rural central Wisconsin, incorpor- 
ating all or parts of seven counties. The Marshfield 
Medical Foundation is the incorpoYated entity which 
was established by GMCHP to exclusively administer 
the plan's Federal programs. 

Under the demonstration, GMCHP is being offered to 
18,000 Medicare beneficiaries through a continuous 
open enrollment. GMCHP has developed special 
marketing methods and materials to accommodate the 
Medicare beneficiary. Numerous meetings of public 
and senior citizen organizations have been held, and a 
special enrollment office has been established at the 
clinic. The plan has also pursued educational activities 
aimed at Medicare beneficiaries. These efforts resulted 
in enrollment of approximately 6,000 Medicare benefi- 
ciaries between April and August 1980. Marshfield's 
second enrollment campaign increased total plan 
members to 7,436, of whom 143 have disenrolled (68 
percent due to death). Marshfield's generally successful 
enrollment was expected, since the demonstration 
includes the major providers in the service area. Total 
enrollment was targeted at 12,000 by March 1981, but 
only 7,929 had enrolled by the end of February. 

GMCHP benefits, which encompass and extend 
Medicare benefits and eliminate deductibles and 
copayments, are financed by a fixed prospective capi- 
tation payment from HCFA combined with an enrollee 
premium. The capitation rate has been derived from an 
actuarial adjustment of the GMCHP community rate 
which reflects the greater per capita utilization of ser- 
vices by the Medicare population. HCFA's capitation 
payment for the first contract period was set at 99 
percent of the Area Prevailing Cost (APC) for non-chronic 
renal disease (CRD) beneficiaries and 98 percent for 
the second contract period, which began October 1, 
1980. A special capitation rate for the CRD benefici- 
aries has been set at 95 percent of the CRD AAPCC for 
both the first and second contract period. 


The Kaiser Permanente medical program is the largest 
non-governmental health care provider in the world. 
The program provides comprehensive care to over 3.6 
million persons in seven geographic areas. Kaiser is 
conducting the Medicare demonstration in the Oregon 
region, which includes five counties and currently 
serves more than 220,000 persons, or about 20 percent 
of the population of the Portland-Vancouver metro- 
politan area. The Oregon region maintains two hospi- 
tals, several ambulatory care facilities, a mental health 


center, and three dental facilities. Kaiser-Portland 
currently has a GPPP (Section 1833) Medicare contract 
under which 15,000 Medicare beneficiaries receive Part 
B services. There are approximately 309,000 Medicare 
beneficiaries in the region. 

Kaiser's demonstration project with HCFA, which 
began in June 1980, has an enrollment of approxi- 
mately 6,000 beneficiaries (4,500 new enrollees) in a 
comprehensive Medicare plan (M-Plan), which provides 
the entire Medicare benefit package as well as some 
benefits not covered or only partially covered under 
Medicare. Eligibility is limited to Medicare beneficiaries 
who have Parts A and B coverage, qualify for Medicare 
aged or disabled coverage, and live in the health plan 
service area. Enrollment is not offered to Medicare 
members who are institutionalized or who qualify for 
ESRD at the time of application, or to Medicare 
members who have Part B coverage only. 

The M-Plan offers Medicare Parts A and B covered 
benefits without deductible or coinsurance limitations. 
The benefit package includes routine physical exami- 
nations, examinations for eyeglasses and most 
immunizations, full coverage for prescribed home 
health care, and outpatient mental health services 
(non-psychiatric). Under the demonstration. Medicare 
pays 95 percent of the AAPCC for the Portland metro- 
politan area, and the Kaiser rate of profit, as reflected 
in the Adjusted Community Rate, is limited to that for 
private enrollees. The difference between 95 percent of 
the AAPCC and the ACR, called savings, will be 
returned to the Medicare beneficiaries as a reduction in 
dues and/or as new services. During the first enrollment, 
enrollees were randomly assigned to two groups: one 
group was offered the special Medicare M-Plan program 
with no monthly dues. The second group was offered the 
M-Plan or a choice of the M-Plan plus three benefit 
options: (a) no dues; (b) eyeglasses, hearing aids, and 
drugs (with co-payment)— $6 dues; (c) dental care and 
dentures— $9.81 dues; and (d) eyeglasses, hearing aids, 
and dental care— $15.81 dues. Kaiser employed a variety 
of marketing approaches, including spot television 
advertising, to identify the most effective enrollment 

InterStudy— A Medicare Multiple Choice 
Program for Minneapolis/St. Paul 

InterStudy is acting as a broker for four HMOs that are 
testing a competitive market model for Medicare 
beneficiaries in the seven counties which compose the 
Minneapolis/St. Paul area. Beginning in May 1981, 
during open enrollment, each of the participating plans 
will offer a low option plan consisting of all Medicare A 
and B services, plus at least one additional service. In 
addition, all HMOs will offer a high option plan with a 
health screen. During open enrollment, all applicants 
who fail the health screen are offered the low option 
plan. During the remainder of the year, those 
individuals who fail the health screen are offered the 
opportunity to enroll during the next open enrollment 

InterStudy is testing a "wise-buyer" public education 
effort to heighten awareness of the choice of plans. 


InterStudy will prepare educational material to provide 
information on the service package of each 
participating plan and discuss its advantages and 
disadvantages. Reimbursement rates are set at 95 
percent of the AAPCC, based on the actuarial category 
into which each enrollee falls. A key hypothesis of the 
project is that competition among the plans for 
Medicare enrollees will cause each to offer as many 
additional benefits as is financially possible. During 
HCFA negotiations with the four HMOs, the following 
issues surfaced: 

• The HMOs were very concerned that they would 
experience unfavorable selection during open enroll- 
ment. They believed that a fixed monthly premium 
would encourage unfavorable selection, particularly 
for IPA model HMOs. 

• During the first year, the HMOs were very cautious 
about offering many additional benefits as part of the 
low option because they had no previous experience 
with the Medicare population. 

• The HMOs believe they should be able to health 
screen during the remainder of the year to counter- 
balance the open enrollment requirement. 

Health Central 

Health Central is a Federally-qualified staff model HMO 
which initiated operations in January 1978. The HMO 
serves the Ingham-Eaton-Clinton county region of 
Michigan, which comprises Lansing and surrounding 
areas. Initially, Health Central anticipated non-Federal 
enrollment levels of 4,000, 10,000, and 18,000 in its first 
three years of operation. Enrollment far exceeded 
those projections and was frozen at approximately 
22,000 in early 1979. 

Health Central signed a cost reimbursement contract 
with HCFA in October 1978 for services to Medicare 
beneficiaries. However, only a handful of Medicare 
beneficiaries have been enrolled under this contract 
because Health Central's service capacity has been 
filled by the unexpectedly high enrollment levels 
described above. It is anticipated tnat the cost reim- 
bursement contract will be replaced by the risk reim- 
bursement contract when the second phase of the 
demonstration begins. 

This demonstration project will involve both Medicare 
and Medicaid beneficiaries. All covered Medicare ser- 
vices will be furnished, as will all Medicaid-covered 
services with the exception of dental benefits. The plan 
will offer additional benefits to the standard benefit 
packages. Approximately 4,000 beneficiaries (2,200 
Medicare and 2,200 Medicaid) are expected to be 
enrolled within the first two years of this part of the 

Phase I (the design) of the demonstration began in 
November 1978. During this phase. Health Central 
encountered serious financial problems which threat- 
ened its continuance. The financial difficulties were 
caused by excessively rapid enrollment, which resulted 
in a loss of management control over utilization. HCFA 

suspended negotiations with Health Central on the 
demonstration during that period. Those problems 
have now been resolved, principally through the finan- 
cial assistance of Blue Cross and Blue Shield of 
Michigan, which obtained control of the HMO in 
August 1979. 

Most of 1980 was spent negotiating an acceptable 
reimbursement risk arrangement, recognizing Health 
Central's minimal financial reserves, as well as 
resolving flaws in the design protocol. Tentative agree- 
ment has been reached on the risk arrangement, and 
the design issues have been resolved. A May 1981 start-up 
of Phase II activities is now anticipated. 

Kitsap Physicians Service 

Kitsap Physicians Service (KPS) of Bremerton, Wash- 
ington, is a non-Federally qualified HMO which was 
established in 1946. It is sponsored by all of the prac- 
ticing physicians in the plan's service areas (three coun- 
ties) and planned to offer Federal beneficiaries addi- 
tional services and reduced cost-sharing incentives for 
enrollment. The plan currently services over 62,000 
people. KPS originally planned to offer 10,000 
Medicare beneficiaries a more comprehensive alterna- 
tive to their existing coverage under Federal law. Under 
the risk contract, rates were to be based on an 
actuarial approach, with the final rate expected to be at 
or below 95 percent of the AAPCC. 

Kitsap noted that hospital days in its area were already 
lower than the national average and believed that 
hospital costs were rising faster than would be 
recognized in a prospective APC. Since Kitsap wanted 
to pay its physicians higher than Medicare reasonable 
charges, but was unwilling to assume it could reduce 
hospital days, it was unable to achieve a rate lower 
than the APC. Kitsap requested a rate 15 percent more 
than the APC. HCFA therefore decided to terminate the 

Blue Shieid of l\/lassacliusetts 

The contract with Blue Shield of Massachusetts, to 
develop and establish a new HMO to be known as the 
Boston Community Health Plan, Inc. (BCHP), expired 
at the end of Phase I on September 14, 1980. The HMO 
proved infeasible due to the conflicting organizational 
priorities of the participants in the venture, as well as 
the inability of Blue Shield to accept a Medicare rate 
which would result in cost savings for HCFA and Medi- 
care beneficiaries. 

The Blue Shield proposal called for an HMO with a 
unique organizational network of providers: seven 
neighborhood health centers and the Primary Care 
Center of Boston City Hospital (BCH) to deliver ambu- 
latory services, one or two other hospitals to offer 
acute care, and other selected specialists and facilities 
to provide the full range of Medicare and Medicaid 
benefits. It was hoped that the project could demon- 
strate a capitation and service delivery model that 

would serve as an alternative to the traditional, costly, 
urban delivery system and which might be replicated in 
other cities. 

Enrollment would have been open to any resident of 
the city; however, the 370,000 persons living in the 
neighborhoods served by the health centers and BCH 
were targeted. Blue Shield initially planned for BCHP 
to enroll Medicare and Medicaid beneficiaries, near 
poor residents, municipal employees, Federal 
employees, and Blue Cross and Blue Shield sub- 
scribers. Marketing was to be handled primarily 
through door-to-door solicitation by neighborhood 

Blue Shield's chief reason for proposing the project 
was to demonstrate its corporate capability of holding 
an HMO license. The city of Boston, which was a sub- 
contractor to Blue Shield, was primarily seeking a 
vehicle for reducing the financial burden of near poor 
residents receiving care at BCH and health centers. It 
eventually became apparent that Blue Shield believed it 
could not accept a rate set at or near 95 percent of the 
county area cost. Blue Shield argued that suburban 
Suffolk County costs were significantly lower than the 
actual costs of inner-city Boston. Because of the 
infeasibility of calculating a more sensitive area cost in 
the near future, HCFA suggested to Blue Shield that 
BCHP receive cost reimbursement for the demonstra- 
tion's first year. In July, however, Blue Shield notified 
HCFA that it would take the State 12 to 14 months to 
review and approve an HMO license involving the 
complex relationship and beneficiary groups proposed 
for BCHP. This development, combined with all the 
other barriers which emerged during the course of 
Phase I, indicated that BCHP was not feasible. 

Appendix B: Institutionalization Data for 
Determination of the AAPCC 

The computation of the AAPCC requires the determina- 
tion of the number of total Medicare beneficiaries and 
HMO enrollees who are institutionalized in the HMO's 
service area. Since the definition of "institution" is con- 
siderably broader than Medicare SNF, institutionaliza- 
tion data are not available from Medicare data sources. 
Three separate data sources were explored for the 
capitation demonstrations: (1) conducting a special 
survey; (2) obtaining the data from the State Medicaid 
agency; and (3) obtaining the data from national 

The survey form which follows was established and 
mailed by HCFA's Office of Research, Demonstrations, 
and Statistics (ORDS) to all the institutions in the 
demonstrations' service area. The form requests the 
Medicare number of all beneficiaries residing in the 
institutions for over 30 days. The survey was used to 
collect institutional data for the Fallon and Marshfield 
projects. If an institution did not respond within 30 
days, ORDS sent a follow-up letter the response 
rate from both survey was approximately 70 percent, 
with most large institutions responding. A review of 
the responses indicated that Medicaid numbers were 


sometimes provided rather than Medicare numbers, 
and some institutions responded that they were no 
Medicare-certified and therefore received no Medicare 
reimbursement. Those facilities sending erroneous 
data were sent a follow-up letter. When the Medicare 
numbers were matched against the HCFA Master 
Beneficiary Records File for the HMO's service area, 
a large percentage of the numbers submitted did not 
match the numbers on the Master Beneficiary Record 
File. Approximately 30 percent (1,008 out of 3,280) 
and 22 percent (448 out of 2,240) of the data submitted 
for the Fallon and Marshfield projects, respectively, did 
not match. 

The special survey appears to be the most reliable but 
time-consuming procedure to obtain data on institu- 
tional populations. There is clearly a problem with the 
reliability of the data submitted, but there is no obvious 
way to improve the accuracy of voluntarily-submitted 
information. It may not be practical if a large number 
of HMOs elect risk reimbursement in the future. It is 
also not clear whether institutions will continue to 
respond on an annual basis to survey requests. 

The optimal procedure for obtaining accurate 
institutional data is to obtain the data directly from the 
State Medicaid program. This procedure was followed 
for the InterStudy and Kaiser demonstrations. Oregon 
was only able to provide statistics on all Medicare eli- 
gibles who were institutionalized, rather than those 
who were institutionalized for over 30 days. These data 
were acceptable to Kaiser, Portland. One limitation of 
this procedure is that the institutions included must be 
reimbursed by Medicaid. This may result in a more 
restricted set than was included in the survey. This 
problem is mitigated because the data on the HMO and 
service area institutional population were obtained 
from the same source. 

Neither conducting a special survey nor requesting the 
institutional data from the State provides a simple and 
rapid procedure for collecting the required data. The 
third alternative that was explored, but not actually 

used for the demonstrations, is to collect the data from 
an existing survey. The following options were 

(1) In 1976, the Public Health Service (PHS) con- 
ducted the Master Facility Inventory, a survey of 
all institutions in the U.S. This survey provided a 
count of institutionalized persons by county, but 
not broken out by Medicare eligibility or age and 
sex. These can only be determined by extrapola- 
tion. The percentage of elderly on Medicare is 
known, so the Medicare institutionalized popula- 
tion can be approximated by using the national 
age-sex distribution obtained from another PHS 
survey, the National Nursing Home Survey. How- 
ever, no comparable procedure could be estab- 
lished for the disabled population. 

(2) In 1977, the Census Bureau conducted a Survey 
of Institutionalized Persons. This survey, which 
consisted of a national stratified sampling of 
institutions, provided an age-sex breakdown for 
the nation. We have obtained the percent of the 
total institutionalized population by age and sex 
from the Census Bureau on a State level. These 
percentages can be applied to data from the 
Master Facility Survey to obtain actual counts. 
The Bureau has indicated that "county estimates 
would be virtually useless and in some cases, im- 
possible to produce (for example, in counties 
where no sample was taken)." As with the first 
option, assumptions must be made concerning th€ 
percentage of the aged who are Medicare eligible. 
We are unaware of any data that provide compa- 
rable estimates of the percentage of institution- 
alized individuals under 65 who are Medicare 

Although the above procedures require some extrapo- 
lation which may not be acceptable for actuarial 
purposes, they merit further exploration and refinement 
if the collection of institutional data is not to become 
excessively burdensome for HCFA. They could also be 
used to determine rough estimates of the AAPCC. 


Form Approved 
0MB No. 66-R1037 
Survey of Medicare Institutionalized Beneficiaries 

Institution Name Type of Institution 

Address Nursing Home 

Person to Contact: , Convalescent Home 

Telephone Number: Rest Home 


Long Term Hospital 

Other (Specify) 

Please list the Name and Medicare Number of all Medicare beneficiaries who are now residing in your institution, and 
have resided there for at least 30 days. 

Name (Medicare Number 










(Please attach additional sheets if necessary) 

This report is authorized by Public Law 92-603. While you are not required to respond, your cooperation is needed to 
make the results of this study comprehensive, accurate, and timely. 

Please Return This Form To: 

Sidney Trieger, Director 

Division of Health Systems and Special Studies 

Office of Demonstrations and Evaluations 

Room 1-E-5 Oak Meadows Building 

6401 Security Boulevard 

Baltimore, Maryland 21235 


*UA GOVERNMENT PRINTING OFFICE:1982 522-027/6037 1-3 

Health Care 


U.S. Department of Health and Human Services 

Richard R. Schweiker, Secretary 

Health Care Financing Administration 

Carolyne K. Davis, Administrator 

Office of Research, Demonstrations, and 


James M. Kaple, Acting Director 

Jean LeMasurier, Director, Program Planning 
and Support 

Research Publications Group 

Karen Peiham O'Steen, Publications Coordinator 
Donna L. Eskow, Writer-editor 
Carol J. Pianalto, Writer-editor 
Alice L. Young, Writer-editor 
Cynthia Dingle, Editorial Assistant 

Send changes of address or requests for this publication to: 

ORDS Publications 

Rm 1E9 Oak Meadows Building 

6340 Security Blvd. 

Baltimore, MD 21235. 



3 ams D0D1E03S a